Hugging Face's logo

nateraw
/
videomae-base-finetuned-ucf101

Video Classification
Transformers
PyTorch
English
videomae
vision
Inference Endpoints
Model card Files Community
1
nateraw commited on
Commit
8fcb3b6
•
1 Parent(s): 94a3bb4

Upload . with huggingface_hub

Browse files
Files changed (11) hide show
  1. README.md +255 -0
  2. config.json +236 -0
  3. eval_results.json +8 -0
  4. optimizer.pt +3 -0
  5. preprocessor_config.json +18 -0
  6. pytorch_model.bin +3 -0
  7. rng_state.pth +3 -0
  8. scaler.pt +3 -0
  9. scheduler.pt +3 -0
  10. trainer_state.json +3426 -0
  11. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: en
3
+ license: mit
4
+ library_name: transformers
5
+ tags:
6
+ - video-classification
7
+ - videomae
8
+ - vision
9
+ ---
10
+
11
+ # Model Card for videomae-base-finetuned-ucf101
12
+
13
+ <!-- Provide a quick summary of what the model is/does. -->
14
+
15
+ # Table of Contents
16
+
17
+ 1. [Model Details](#model-details)
18
+ 2. [Uses](#uses)
19
+ 3. [Bias, Risks, and Limitations](#bias-risks-and-limitations)
20
+ 4. [Training Details](#training-details)
21
+ 5. [Evaluation](#evaluation)
22
+ 6. [Model Examination](#model-examination-optional)
23
+ 7. [Environmental Impact](#environmental-impact)
24
+ 8. [Technical Specifications](#technical-specifications-optional)
25
+ 9. [Citation](#citation-optional)
26
+ 10. [Glossary](#glossary-optional)
27
+ 11. [More Information](#more-information-optional)
28
+ 12. [Model Card Authors](#model-card-authors-optional)
29
+ 13. [Model Card Contact](#model-card-contact)
30
+ 14. [How To Get Started With the Model](#how-to-get-started-with-the-model)
31
+
32
+
33
+ # Model Details
34
+
35
+ ## Model Description
36
+
37
+ <!-- Provide a longer summary of what this model is. -->
38
+
39
+ VideoMAE Base model fine tuned on UCF101
40
+
41
+ - **Developed by:** [@nateraw](https://huggingface.co/nateraw)
42
+ - **Shared by [optional]:** [More Information Needed]
43
+ - **Model type:** fine-tuned
44
+ - **Language(s) (NLP):** en
45
+ - **License:** mit
46
+ - **Related Models [optional]:** [More Information Needed]
47
+ - **Parent Model [optional]:** [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base)
48
+ - **Resources for more information:** [More Information Needed]
49
+
50
+ # Uses
51
+
52
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
53
+
54
+ ## Direct Use
55
+
56
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
57
+
58
+ This model can be used for Video Action Recognition
59
+
60
+ ## Downstream Use [optional]
61
+
62
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
63
+
64
+ [More Information Needed]
65
+
66
+ ## Out-of-Scope Use
67
+
68
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
69
+
70
+ [More Information Needed]
71
+
72
+ # Bias, Risks, and Limitations
73
+
74
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
75
+
76
+ [More Information Needed]
77
+
78
+ ## Recommendations
79
+
80
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
81
+
82
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recomendations.
83
+
84
+ # Training Details
85
+
86
+ ## Training Data
87
+
88
+ <!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
89
+
90
+ [More Information Needed]
91
+
92
+ ## Training Procedure [optional]
93
+
94
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
95
+
96
+ ### Preprocessing
97
+
98
+ We sampled clips from the videos of 64 frames, then took a uniform sample of those frames to get 16 frame inputs for the model. During training, we used PyTorchVideo's [`MixVideo`](https://github.com/facebookresearch/pytorchvideo/blob/main/pytorchvideo/transforms/mix.py) to apply mixup/cutmix.
99
+
100
+ ### Speeds, Sizes, Times
101
+
102
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
103
+
104
+ [More Information Needed]
105
+
106
+ # Evaluation
107
+
108
+ <!-- This section describes the evaluation protocols and provides the results. -->
109
+
110
+ ## Testing Data, Factors & Metrics
111
+
112
+ ### Testing Data
113
+
114
+ <!-- This should link to a Data Card if possible. -->
115
+
116
+ [More Information Needed]
117
+
118
+ ### Factors
119
+
120
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
121
+
122
+ [More Information Needed]
123
+
124
+ ### Metrics
125
+
126
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
127
+
128
+ [More Information Needed]
129
+
130
+ ## Results
131
+
132
+ We only trained/evaluated one fold from the UCF101 annotations. Unlike in the VideoMAE paper, we did not perform inference over multiple crops/segments of validation videos, so the results are likely slightly lower than what you would get if you did that too.
133
+ - Eval Accuracy: 0.758209764957428
134
+ - Eval Accuracy Top 5: 0.8983050584793091
135
+
136
+ # Model Examination [optional]
137
+
138
+ <!-- Relevant interpretability work for the model goes here -->
139
+
140
+ [More Information Needed]
141
+
142
+ # Environmental Impact
143
+
144
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
145
+
146
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
147
+
148
+ - **Hardware Type:** [More Information Needed]
149
+ - **Hours used:** [More Information Needed]
150
+ - **Cloud Provider:** [More Information Needed]
151
+ - **Compute Region:** [More Information Needed]
152
+ - **Carbon Emitted:** [More Information Needed]
153
+
154
+ # Technical Specifications [optional]
155
+
156
+ ## Model Architecture and Objective
157
+
158
+ [More Information Needed]
159
+
160
+ ## Compute Infrastructure
161
+
162
+ [More Information Needed]
163
+
164
+ ### Hardware
165
+
166
+ [More Information Needed]
167
+
168
+ ### Software
169
+
170
+ [More Information Needed]
171
+
172
+ # Citation [optional]
173
+
174
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
175
+
176
+ **BibTeX:**
177
+
178
+ [More Information Needed]
179
+
180
+ **APA:**
181
+
182
+ [More Information Needed]
183
+
184
+ # Glossary [optional]
185
+
186
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
187
+
188
+ [More Information Needed]
189
+
190
+ # More Information [optional]
191
+
192
+ [More Information Needed]
193
+
194
+ # Model Card Authors [optional]
195
+
196
+ [@nateraw](https://huggingface.co/nateraw)
197
+
198
+ # Model Card Contact
199
+
200
+ [@nateraw](https://huggingface.co/nateraw)
201
+
202
+ # How to Get Started with the Model
203
+
204
+ Use the code below to get started with the model.
205
+
206
+ <details>
207
+ <summary> Click to expand </summary>
208
+
209
+ ```python
210
+ from decord import VideoReader, cpu
211
+ import torch
212
+ import numpy as np
213
+
214
+ from transformers import VideoMAEFeatureExtractor, VideoMAEForVideoClassification
215
+ from huggingface_hub import hf_hub_download
216
+
217
+ np.random.seed(0)
218
+
219
+
220
+ def sample_frame_indices(clip_len, frame_sample_rate, seg_len):
221
+ converted_len = int(clip_len * frame_sample_rate)
222
+ end_idx = np.random.randint(converted_len, seg_len)
223
+ start_idx = end_idx - converted_len
224
+ indices = np.linspace(start_idx, end_idx, num=clip_len)
225
+ indices = np.clip(indices, start_idx, end_idx - 1).astype(np.int64)
226
+ return indices
227
+
228
+
229
+ # video clip consists of 300 frames (10 seconds at 30 FPS)
230
+ file_path = hf_hub_download(
231
+ repo_id="nateraw/dino-clips", filename="archery.mp4", repo_type="space"
232
+ )
233
+ videoreader = VideoReader(file_path, num_threads=1, ctx=cpu(0))
234
+
235
+ # sample 16 frames
236
+ videoreader.seek(0)
237
+ indices = sample_frame_indices(clip_len=16, frame_sample_rate=4, seg_len=len(videoreader))
238
+ video = videoreader.get_batch(indices).asnumpy()
239
+
240
+ feature_extractor = VideoMAEFeatureExtractor.from_pretrained("nateraw/videomae-base-finetuned-ucf101")
241
+ model = VideoMAEForVideoClassification.from_pretrained("nateraw/videomae-base-finetuned-ucf101")
242
+
243
+ inputs = feature_extractor(list(video), return_tensors="pt")
244
+
245
+ with torch.no_grad():
246
+ outputs = model(**inputs)
247
+ logits = outputs.logits
248
+
249
+ # model predicts one of the 101 UCF101 classes
250
+ predicted_label = logits.argmax(-1).item()
251
+ print(model.config.id2label[predicted_label])
252
+ ```
253
+
254
+
255
+ </details>
config.json ADDED
@@ -0,0 +1,236 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "MCG-NJU/videomae-base",
3
+ "architectures": [
4
+ "VideoMAEForVideoClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "decoder_hidden_size": 384,
8
+ "decoder_intermediate_size": 1536,
9
+ "decoder_num_attention_heads": 6,
10
+ "decoder_num_hidden_layers": 4,
11
+ "hidden_act": "gelu",
12
+ "hidden_dropout_prob": 0.0,
13
+ "hidden_size": 768,
14
+ "id2label": {
15
+ "0": "ApplyEyeMakeup",
16
+ "1": "ApplyLipstick",
17
+ "2": "Archery",
18
+ "3": "BabyCrawling",
19
+ "4": "BalanceBeam",
20
+ "5": "BandMarching",
21
+ "6": "BaseballPitch",
22
+ "7": "Basketball",
23
+ "8": "BasketballDunk",
24
+ "9": "BenchPress",
25
+ "10": "Biking",
26
+ "11": "Billiards",
27
+ "12": "BlowDryHair",
28
+ "13": "BlowingCandles",
29
+ "14": "BodyWeightSquats",
30
+ "15": "Bowling",
31
+ "16": "BoxingPunchingBag",
32
+ "17": "BoxingSpeedBag",
33
+ "18": "BreastStroke",
34
+ "19": "BrushingTeeth",
35
+ "20": "CleanAndJerk",
36
+ "21": "CliffDiving",
37
+ "22": "CricketBowling",
38
+ "23": "CricketShot",
39
+ "24": "CuttingInKitchen",
40
+ "25": "Diving",
41
+ "26": "Drumming",
42
+ "27": "Fencing",
43
+ "28": "FieldHockeyPenalty",
44
+ "29": "FloorGymnastics",
45
+ "30": "FrisbeeCatch",
46
+ "31": "FrontCrawl",
47
+ "32": "GolfSwing",
48
+ "33": "Haircut",
49
+ "34": "HammerThrow",
50
+ "35": "Hammering",
51
+ "36": "HandstandPushups",
52
+ "37": "HandstandWalking",
53
+ "38": "HeadMassage",
54
+ "39": "HighJump",
55
+ "40": "HorseRace",
56
+ "41": "HorseRiding",
57
+ "42": "HulaHoop",
58
+ "43": "IceDancing",
59
+ "44": "JavelinThrow",
60
+ "45": "JugglingBalls",
61
+ "46": "JumpRope",
62
+ "47": "JumpingJack",
63
+ "48": "Kayaking",
64
+ "49": "Knitting",
65
+ "50": "LongJump",
66
+ "51": "Lunges",
67
+ "52": "MilitaryParade",
68
+ "53": "Mixing",
69
+ "54": "MoppingFloor",
70
+ "55": "Nunchucks",
71
+ "56": "ParallelBars",
72
+ "57": "PizzaTossing",
73
+ "58": "PlayingCello",
74
+ "59": "PlayingDaf",
75
+ "60": "PlayingDhol",
76
+ "61": "PlayingFlute",
77
+ "62": "PlayingGuitar",
78
+ "63": "PlayingPiano",
79
+ "64": "PlayingSitar",
80
+ "65": "PlayingTabla",
81
+ "66": "PlayingViolin",
82
+ "67": "PoleVault",
83
+ "68": "PommelHorse",
84
+ "69": "PullUps",
85
+ "70": "Punch",
86
+ "71": "PushUps",
87
+ "72": "Rafting",
88
+ "73": "RockClimbingIndoor",
89
+ "74": "RopeClimbing",
90
+ "75": "Rowing",
91
+ "76": "SalsaSpin",
92
+ "77": "ShavingBeard",
93
+ "78": "Shotput",
94
+ "79": "SkateBoarding",
95
+ "80": "Skiing",
96
+ "81": "Skijet",
97
+ "82": "SkyDiving",
98
+ "83": "SoccerJuggling",
99
+ "84": "SoccerPenalty",
100
+ "85": "StillRings",
101
+ "86": "SumoWrestling",
102
+ "87": "Surfing",
103
+ "88": "Swing",
104
+ "89": "TableTennisShot",
105
+ "90": "TaiChi",
106
+ "91": "TennisSwing",
107
+ "92": "ThrowDiscus",
108
+ "93": "TrampolineJumping",
109
+ "94": "Typing",
110
+ "95": "UnevenBars",
111
+ "96": "VolleyballSpiking",
112
+ "97": "WalkingWithDog",
113
+ "98": "WallPushups",
114
+ "99": "WritingOnBoard",
115
+ "100": "YoYo"
116
+ },
117
+ "image_size": 224,
118
+ "initializer_range": 0.02,
119
+ "intermediate_size": 3072,
120
+ "label2id": {
121
+ "ApplyEyeMakeup": 0,
122
+ "ApplyLipstick": 1,
123
+ "Archery": 2,
124
+ "BabyCrawling": 3,
125
+ "BalanceBeam": 4,
126
+ "BandMarching": 5,
127
+ "BaseballPitch": 6,
128
+ "Basketball": 7,
129
+ "BasketballDunk": 8,
130
+ "BenchPress": 9,
131
+ "Biking": 10,
132
+ "Billiards": 11,
133
+ "BlowDryHair": 12,
134
+ "BlowingCandles": 13,
135
+ "BodyWeightSquats": 14,
136
+ "Bowling": 15,
137
+ "BoxingPunchingBag": 16,
138
+ "BoxingSpeedBag": 17,
139
+ "BreastStroke": 18,
140
+ "BrushingTeeth": 19,
141
+ "CleanAndJerk": 20,
142
+ "CliffDiving": 21,
143
+ "CricketBowling": 22,
144
+ "CricketShot": 23,
145
+ "CuttingInKitchen": 24,
146
+ "Diving": 25,
147
+ "Drumming": 26,
148
+ "Fencing": 27,
149
+ "FieldHockeyPenalty": 28,
150
+ "FloorGymnastics": 29,
151
+ "FrisbeeCatch": 30,
152
+ "FrontCrawl": 31,
153
+ "GolfSwing": 32,
154
+ "Haircut": 33,
155
+ "HammerThrow": 34,
156
+ "Hammering": 35,
157
+ "HandstandPushups": 36,
158
+ "HandstandWalking": 37,
159
+ "HeadMassage": 38,
160
+ "HighJump": 39,
161
+ "HorseRace": 40,
162
+ "HorseRiding": 41,
163
+ "HulaHoop": 42,
164
+ "IceDancing": 43,
165
+ "JavelinThrow": 44,
166
+ "JugglingBalls": 45,
167
+ "JumpRope": 46,
168
+ "JumpingJack": 47,
169
+ "Kayaking": 48,
170
+ "Knitting": 49,
171
+ "LongJump": 50,
172
+ "Lunges": 51,
173
+ "MilitaryParade": 52,
174
+ "Mixing": 53,
175
+ "MoppingFloor": 54,
176
+ "Nunchucks": 55,
177
+ "ParallelBars": 56,
178
+ "PizzaTossing": 57,
179
+ "PlayingCello": 58,
180
+ "PlayingDaf": 59,
181
+ "PlayingDhol": 60,
182
+ "PlayingFlute": 61,
183
+ "PlayingGuitar": 62,
184
+ "PlayingPiano": 63,
185
+ "PlayingSitar": 64,
186
+ "PlayingTabla": 65,
187
+ "PlayingViolin": 66,
188
+ "PoleVault": 67,
189
+ "PommelHorse": 68,
190
+ "PullUps": 69,
191
+ "Punch": 70,
192
+ "PushUps": 71,
193
+ "Rafting": 72,
194
+ "RockClimbingIndoor": 73,
195
+ "RopeClimbing": 74,
196
+ "Rowing": 75,
197
+ "SalsaSpin": 76,
198
+ "ShavingBeard": 77,
199
+ "Shotput": 78,
200
+ "SkateBoarding": 79,
201
+ "Skiing": 80,
202
+ "Skijet": 81,
203
+ "SkyDiving": 82,
204
+ "SoccerJuggling": 83,
205
+ "SoccerPenalty": 84,
206
+ "StillRings": 85,
207
+ "SumoWrestling": 86,
208
+ "Surfing": 87,
209
+ "Swing": 88,
210
+ "TableTennisShot": 89,
211
+ "TaiChi": 90,
212
+ "TennisSwing": 91,
213
+ "ThrowDiscus": 92,
214
+ "TrampolineJumping": 93,
215
+ "Typing": 94,
216
+ "UnevenBars": 95,
217
+ "VolleyballSpiking": 96,
218
+ "WalkingWithDog": 97,
219
+ "WallPushups": 98,
220
+ "WritingOnBoard": 99,
221
+ "YoYo": 100
222
+ },
223
+ "layer_norm_eps": 1e-12,
224
+ "model_type": "videomae",
225
+ "norm_pix_loss": true,
226
+ "num_attention_heads": 12,
227
+ "num_channels": 3,
228
+ "num_frames": 16,
229
+ "num_hidden_layers": 12,
230
+ "patch_size": 16,
231
+ "qkv_bias": true,
232
+ "torch_dtype": "float32",
233
+ "transformers_version": "4.24.0",
234
+ "tubelet_size": 2,
235
+ "use_mean_pooling": false
236
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "eval_accuracy": 0.758209764957428,
3
+ "eval_accuracy_top5": 0.8983050584793091,
4
+ "eval_loss": 1.1774553060531616,
5
+ "eval_runtime": 110.0091,
6
+ "eval_samples_per_second": 34.388,
7
+ "eval_steps_per_second": 4.3
8
+ }
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03d088f0159e7009c1c171dc43b7432714239a5cb68671293de7999c19fd9be0
3
+ size 690548101
preprocessor_config.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_center_crop": true,
3
+ "do_normalize": true,
4
+ "do_resize": true,
5
+ "feature_extractor_type": "VideoMAEFeatureExtractor",
6
+ "image_mean": [
7
+ 0.485,
8
+ 0.456,
9
+ 0.406
10
+ ],
11
+ "image_std": [
12
+ 0.229,
13
+ 0.224,
14
+ 0.225
15
+ ],
16
+ "resample": 2,
17
+ "size": 224
18
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c0da29a2e7835c0d9d95e0ea5b1da36e07c6c5bd9996137b3bfdf0834e525d9
3
+ size 345283445
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8868ab8b4428d8f3400d98c910857a7d7c4325ccece5968faddb26b6803c920
3
+ size 14639
scaler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:367928c873a79d90abcb66db0f8b320eea7346dc9bc779b4e7963dfc82cf2ada
3
+ size 557
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:610e42719ee26d38bf25ce431eb179cdfbae56bdb396e5979680d0a22ea4502b
3
+ size 627
trainer_state.json ADDED
@@ -0,0 +1,3426 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.7513241767883301,
3
+ "best_model_checkpoint": "videomae-base-finetuned-ucf101-nomixup/checkpoint-5500",
4
+ "epoch": 36.91275167785235,
5
+ "global_step": 5500,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.07,
12
+ "learning_rate": 6.7114093959731546e-06,
13
+ "loss": 4.685,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.13,
18
+ "learning_rate": 1.3422818791946309e-05,
19
+ "loss": 4.6973,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.2,
24
+ "learning_rate": 2.013422818791946e-05,
25
+ "loss": 4.6677,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.27,
30
+ "learning_rate": 2.6845637583892618e-05,
31
+ "loss": 4.6252,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.34,
36
+ "learning_rate": 3.3557046979865775e-05,
37
+ "loss": 4.6242,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.4,
42
+ "learning_rate": 4.026845637583892e-05,
43
+ "loss": 4.5581,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.47,
48
+ "learning_rate": 4.697986577181208e-05,
49
+ "loss": 4.5285,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.54,
54
+ "learning_rate": 5.3691275167785237e-05,
55
+ "loss": 4.5188,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.6,
60
+ "learning_rate": 6.040268456375839e-05,
61
+ "loss": 4.431,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.67,
66
+ "learning_rate": 6.711409395973155e-05,
67
+ "loss": 4.3994,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.74,
72
+ "learning_rate": 7.38255033557047e-05,
73
+ "loss": 4.3375,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.81,
78
+ "learning_rate": 8.053691275167784e-05,
79
+ "loss": 4.2659,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.87,
84
+ "learning_rate": 8.7248322147651e-05,
85
+ "loss": 4.1503,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.94,
90
+ "learning_rate": 9.395973154362417e-05,
91
+ "loss": 4.0464,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 1.01,
96
+ "learning_rate": 0.00010067114093959731,
97
+ "loss": 3.9801,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 1.07,
102
+ "learning_rate": 0.00010738255033557047,
103
+ "loss": 3.9336,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 1.14,
108
+ "learning_rate": 0.00011409395973154363,
109
+ "loss": 3.8316,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 1.21,
114
+ "learning_rate": 0.00012080536912751678,
115
+ "loss": 3.7867,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 1.28,
120
+ "learning_rate": 0.00012751677852348994,
121
+ "loss": 3.6977,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 1.34,
126
+ "learning_rate": 0.0001342281879194631,
127
+ "loss": 3.5853,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 1.41,
132
+ "learning_rate": 0.00014093959731543624,
133
+ "loss": 3.4901,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 1.48,
138
+ "learning_rate": 0.0001476510067114094,
139
+ "loss": 3.4679,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 1.54,
144
+ "learning_rate": 0.00015436241610738256,
145
+ "loss": 3.4984,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 1.61,
150
+ "learning_rate": 0.0001610738255033557,
151
+ "loss": 3.3598,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 1.68,
156
+ "learning_rate": 0.00016778523489932888,
157
+ "loss": 3.2679,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 1.74,
162
+ "learning_rate": 0.000174496644295302,
163
+ "loss": 3.2981,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 1.81,
168
+ "learning_rate": 0.00018120805369127517,
169
+ "loss": 3.2664,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 1.88,
174
+ "learning_rate": 0.00018791946308724833,
175
+ "loss": 3.2929,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 1.95,
180
+ "learning_rate": 0.00019463087248322146,
181
+ "loss": 3.2039,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 2.01,
186
+ "learning_rate": 0.00020134228187919463,
187
+ "loss": 3.1763,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 2.08,
192
+ "learning_rate": 0.00020805369127516779,
193
+ "loss": 3.0554,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 2.15,
198
+ "learning_rate": 0.00021476510067114095,
199
+ "loss": 2.9937,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 2.21,
204
+ "learning_rate": 0.0002214765100671141,
205
+ "loss": 3.0402,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 2.28,
210
+ "learning_rate": 0.00022818791946308727,
211
+ "loss": 3.0012,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 2.35,
216
+ "learning_rate": 0.0002348993288590604,
217
+ "loss": 2.9892,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 2.42,
222
+ "learning_rate": 0.00024161073825503356,
223
+ "loss": 3.0453,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 2.48,
228
+ "learning_rate": 0.0002483221476510067,
229
+ "loss": 2.9497,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 2.55,
234
+ "learning_rate": 0.0002550335570469799,
235
+ "loss": 2.8924,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 2.62,
240
+ "learning_rate": 0.000261744966442953,
241
+ "loss": 2.9597,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 2.68,
246
+ "learning_rate": 0.0002684563758389262,
247
+ "loss": 3.0828,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 2.75,
252
+ "learning_rate": 0.00027516778523489934,
253
+ "loss": 3.0353,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 2.82,
258
+ "learning_rate": 0.00028187919463087247,
259
+ "loss": 2.8786,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 2.89,
264
+ "learning_rate": 0.00028859060402684566,
265
+ "loss": 2.9322,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 2.95,
270
+ "learning_rate": 0.0002953020134228188,
271
+ "loss": 2.7467,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 3.02,
276
+ "learning_rate": 0.000302013422818792,
277
+ "loss": 2.8574,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 3.09,
282
+ "learning_rate": 0.0003087248322147651,
283
+ "loss": 2.7929,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 3.15,
288
+ "learning_rate": 0.00031543624161073825,
289
+ "loss": 2.8987,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 3.22,
294
+ "learning_rate": 0.0003221476510067114,
295
+ "loss": 2.8908,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 3.29,
300
+ "learning_rate": 0.0003288590604026846,
301
+ "loss": 2.6509,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 3.36,
306
+ "learning_rate": 0.00033557046979865775,
307
+ "loss": 2.8576,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 3.36,
312
+ "eval_accuracy": 0.5911017060279846,
313
+ "eval_accuracy_top5": 0.727224588394165,
314
+ "eval_loss": 1.8057622909545898,
315
+ "eval_runtime": 104.4766,
316
+ "eval_samples_per_second": 36.209,
317
+ "eval_steps_per_second": 4.527,
318
+ "step": 500
319
+ },
320
+ {
321
+ "epoch": 3.42,
322
+ "learning_rate": 0.0003422818791946309,
323
+ "loss": 2.7528,
324
+ "step": 510
325
+ },
326
+ {
327
+ "epoch": 3.49,
328
+ "learning_rate": 0.000348993288590604,
329
+ "loss": 2.7427,
330
+ "step": 520
331
+ },
332
+ {
333
+ "epoch": 3.56,
334
+ "learning_rate": 0.00035570469798657715,
335
+ "loss": 2.8489,
336
+ "step": 530
337
+ },
338
+ {
339
+ "epoch": 3.62,
340
+ "learning_rate": 0.00036241610738255034,
341
+ "loss": 2.8167,
342
+ "step": 540
343
+ },
344
+ {
345
+ "epoch": 3.69,
346
+ "learning_rate": 0.00036912751677852353,
347
+ "loss": 2.7375,
348
+ "step": 550
349
+ },
350
+ {
351
+ "epoch": 3.76,
352
+ "learning_rate": 0.00037583892617449666,
353
+ "loss": 2.8108,
354
+ "step": 560
355
+ },
356
+ {
357
+ "epoch": 3.83,
358
+ "learning_rate": 0.0003825503355704698,
359
+ "loss": 2.8261,
360
+ "step": 570
361
+ },
362
+ {
363
+ "epoch": 3.89,
364
+ "learning_rate": 0.00038926174496644293,
365
+ "loss": 2.7504,
366
+ "step": 580
367
+ },
368
+ {
369
+ "epoch": 3.96,
370
+ "learning_rate": 0.0003959731543624161,
371
+ "loss": 2.802,
372
+ "step": 590
373
+ },
374
+ {
375
+ "epoch": 4.03,
376
+ "learning_rate": 0.00040268456375838925,
377
+ "loss": 2.7507,
378
+ "step": 600
379
+ },
380
+ {
381
+ "epoch": 4.09,
382
+ "learning_rate": 0.00040939597315436244,
383
+ "loss": 2.8335,
384
+ "step": 610
385
+ },
386
+ {
387
+ "epoch": 4.16,
388
+ "learning_rate": 0.00041610738255033557,
389
+ "loss": 2.6994,
390
+ "step": 620
391
+ },
392
+ {
393
+ "epoch": 4.23,
394
+ "learning_rate": 0.00042281879194630876,
395
+ "loss": 2.8312,
396
+ "step": 630
397
+ },
398
+ {
399
+ "epoch": 4.3,
400
+ "learning_rate": 0.0004295302013422819,
401
+ "loss": 2.8237,
402
+ "step": 640
403
+ },
404
+ {
405
+ "epoch": 4.36,
406
+ "learning_rate": 0.000436241610738255,
407
+ "loss": 2.849,
408
+ "step": 650
409
+ },
410
+ {
411
+ "epoch": 4.43,
412
+ "learning_rate": 0.0004429530201342282,
413
+ "loss": 2.8233,
414
+ "step": 660
415
+ },
416
+ {
417
+ "epoch": 4.5,
418
+ "learning_rate": 0.00044966442953020135,
419
+ "loss": 2.7256,
420
+ "step": 670
421
+ },
422
+ {
423
+ "epoch": 4.56,
424
+ "learning_rate": 0.00045637583892617453,
425
+ "loss": 2.8312,
426
+ "step": 680
427
+ },
428
+ {
429
+ "epoch": 4.63,
430
+ "learning_rate": 0.00046308724832214767,
431
+ "loss": 2.7241,
432
+ "step": 690
433
+ },
434
+ {
435
+ "epoch": 4.7,
436
+ "learning_rate": 0.0004697986577181208,
437
+ "loss": 2.6984,
438
+ "step": 700
439
+ },
440
+ {
441
+ "epoch": 4.77,
442
+ "learning_rate": 0.00047651006711409394,
443
+ "loss": 2.7826,
444
+ "step": 710
445
+ },
446
+ {
447
+ "epoch": 4.83,
448
+ "learning_rate": 0.0004832214765100671,
449
+ "loss": 2.781,
450
+ "step": 720
451
+ },
452
+ {
453
+ "epoch": 4.9,
454
+ "learning_rate": 0.0004899328859060403,
455
+ "loss": 2.7601,
456
+ "step": 730
457
+ },
458
+ {
459
+ "epoch": 4.97,
460
+ "learning_rate": 0.0004966442953020134,
461
+ "loss": 2.744,
462
+ "step": 740
463
+ },
464
+ {
465
+ "epoch": 5.03,
466
+ "learning_rate": 0.0004999988659275162,
467
+ "loss": 2.8382,
468
+ "step": 750
469
+ },
470
+ {
471
+ "epoch": 5.1,
472
+ "learning_rate": 0.0004999897934093798,
473
+ "loss": 2.8987,
474
+ "step": 760
475
+ },
476
+ {
477
+ "epoch": 5.17,
478
+ "learning_rate": 0.0004999716487023506,
479
+ "loss": 2.7493,
480
+ "step": 770
481
+ },
482
+ {
483
+ "epoch": 5.23,
484
+ "learning_rate": 0.0004999444324649045,
485
+ "loss": 2.8423,
486
+ "step": 780
487
+ },
488
+ {
489
+ "epoch": 5.3,
490
+ "learning_rate": 0.0004999081456847252,
491
+ "loss": 2.8108,
492
+ "step": 790
493
+ },
494
+ {
495
+ "epoch": 5.37,
496
+ "learning_rate": 0.0004998627896786686,
497
+ "loss": 2.7675,
498
+ "step": 800
499
+ },
500
+ {
501
+ "epoch": 5.44,
502
+ "learning_rate": 0.0004998083660927148,
503
+ "loss": 2.5845,
504
+ "step": 810
505
+ },
506
+ {
507
+ "epoch": 5.5,
508
+ "learning_rate": 0.0004997448769019087,
509
+ "loss": 2.7445,
510
+ "step": 820
511
+ },
512
+ {
513
+ "epoch": 5.57,
514
+ "learning_rate": 0.0004996723244102881,
515
+ "loss": 2.6608,
516
+ "step": 830
517
+ },
518
+ {
519
+ "epoch": 5.64,
520
+ "learning_rate": 0.0004995907112508002,
521
+ "loss": 2.6423,
522
+ "step": 840
523
+ },
524
+ {
525
+ "epoch": 5.7,
526
+ "learning_rate": 0.0004995000403852056,
527
+ "loss": 2.706,
528
+ "step": 850
529
+ },
530
+ {
531
+ "epoch": 5.77,
532
+ "learning_rate": 0.0004994003151039719,
533
+ "loss": 2.7687,
534
+ "step": 860
535
+ },
536
+ {
537
+ "epoch": 5.84,
538
+ "learning_rate": 0.000499291539026153,
539
+ "loss": 2.761,
540
+ "step": 870
541
+ },
542
+ {
543
+ "epoch": 5.91,
544
+ "learning_rate": 0.0004991737160992588,
545
+ "loss": 2.6858,
546
+ "step": 880
547
+ },
548
+ {
549
+ "epoch": 5.97,
550
+ "learning_rate": 0.000499046850599111,
551
+ "loss": 2.7172,
552
+ "step": 890
553
+ },
554
+ {
555
+ "epoch": 6.04,
556
+ "learning_rate": 0.0004989109471296887,
557
+ "loss": 2.8251,
558
+ "step": 900
559
+ },
560
+ {
561
+ "epoch": 6.11,
562
+ "learning_rate": 0.000498766010622961,
563
+ "loss": 2.6963,
564
+ "step": 910
565
+ },
566
+ {
567
+ "epoch": 6.17,
568
+ "learning_rate": 0.0004986120463387084,
569
+ "loss": 2.4729,
570
+ "step": 920
571
+ },
572
+ {
573
+ "epoch": 6.24,
574
+ "learning_rate": 0.000498449059864331,
575
+ "loss": 2.7372,
576
+ "step": 930
577
+ },
578
+ {
579
+ "epoch": 6.31,
580
+ "learning_rate": 0.0004982770571146468,
581
+ "loss": 2.8074,
582
+ "step": 940
583
+ },
584
+ {
585
+ "epoch": 6.38,
586
+ "learning_rate": 0.0004980960443316762,
587
+ "loss": 2.7279,
588
+ "step": 950
589
+ },
590
+ {
591
+ "epoch": 6.44,
592
+ "learning_rate": 0.0004979060280844162,
593
+ "loss": 2.6071,
594
+ "step": 960
595
+ },
596
+ {
597
+ "epoch": 6.51,
598
+ "learning_rate": 0.0004977070152686013,
599
+ "loss": 2.7364,
600
+ "step": 970
601
+ },
602
+ {
603
+ "epoch": 6.58,
604
+ "learning_rate": 0.000497499013106454,
605
+ "loss": 2.6095,
606
+ "step": 980
607
+ },
608
+ {
609
+ "epoch": 6.64,
610
+ "learning_rate": 0.0004972820291464219,
611
+ "loss": 2.574,
612
+ "step": 990
613
+ },
614
+ {
615
+ "epoch": 6.71,
616
+ "learning_rate": 0.0004970560712629044,
617
+ "loss": 2.6651,
618
+ "step": 1000
619
+ },
620
+ {
621
+ "epoch": 6.71,
622
+ "eval_accuracy": 0.5632944703102112,
623
+ "eval_accuracy_top5": 0.7002118825912476,
624
+ "eval_loss": 1.7887717485427856,
625
+ "eval_runtime": 1057.27,
626
+ "eval_samples_per_second": 3.578,
627
+ "eval_steps_per_second": 0.447,
628
+ "step": 1000
629
+ },
630
+ {
631
+ "epoch": 6.78,
632
+ "learning_rate": 0.0004968211476559667,
633
+ "loss": 2.7145,
634
+ "step": 1010
635
+ },
636
+ {
637
+ "epoch": 6.85,
638
+ "learning_rate": 0.0004965772668510425,
639
+ "loss": 2.6658,
640
+ "step": 1020
641
+ },
642
+ {
643
+ "epoch": 6.91,
644
+ "learning_rate": 0.0004963244376986235,
645
+ "loss": 2.7473,
646
+ "step": 1030
647
+ },
648
+ {
649
+ "epoch": 6.98,
650
+ "learning_rate": 0.0004960626693739401,
651
+ "loss": 2.8268,
652
+ "step": 1040
653
+ },
654
+ {
655
+ "epoch": 7.05,
656
+ "learning_rate": 0.0004957919713766266,
657
+ "loss": 2.5706,
658
+ "step": 1050
659
+ },
660
+ {
661
+ "epoch": 7.11,
662
+ "learning_rate": 0.0004955123535303775,
663
+ "loss": 2.645,
664
+ "step": 1060
665
+ },
666
+ {
667
+ "epoch": 7.18,
668
+ "learning_rate": 0.0004952238259825911,
669
+ "loss": 2.68,
670
+ "step": 1070
671
+ },
672
+ {
673
+ "epoch": 7.25,
674
+ "learning_rate": 0.0004949263992040001,
675
+ "loss": 2.6571,
676
+ "step": 1080
677
+ },
678
+ {
679
+ "epoch": 7.32,
680
+ "learning_rate": 0.0004946200839882932,
681
+ "loss": 2.4641,
682
+ "step": 1090
683
+ },
684
+ {
685
+ "epoch": 7.38,
686
+ "learning_rate": 0.000494304891451722,
687
+ "loss": 2.6196,
688
+ "step": 1100
689
+ },
690
+ {
691
+ "epoch": 7.45,
692
+ "learning_rate": 0.0004939808330326985,
693
+ "loss": 2.5998,
694
+ "step": 1110
695
+ },
696
+ {
697
+ "epoch": 7.52,
698
+ "learning_rate": 0.0004936479204913795,
699
+ "loss": 2.6489,
700
+ "step": 1120
701
+ },
702
+ {
703
+ "epoch": 7.58,
704
+ "learning_rate": 0.0004933061659092401,
705
+ "loss": 2.6613,
706
+ "step": 1130
707
+ },
708
+ {
709
+ "epoch": 7.65,
710
+ "learning_rate": 0.0004929555816886352,
711
+ "loss": 2.533,
712
+ "step": 1140
713
+ },
714
+ {
715
+ "epoch": 7.72,
716
+ "learning_rate": 0.0004925961805523493,
717
+ "loss": 2.6853,
718
+ "step": 1150
719
+ },
720
+ {
721
+ "epoch": 7.79,
722
+ "learning_rate": 0.0004922279755431349,
723
+ "loss": 2.6887,
724
+ "step": 1160
725
+ },
726
+ {
727
+ "epoch": 7.85,
728
+ "learning_rate": 0.0004918509800232392,
729
+ "loss": 2.6397,
730
+ "step": 1170
731
+ },
732
+ {
733
+ "epoch": 7.92,
734
+ "learning_rate": 0.0004914652076739187,
735
+ "loss": 2.4045,
736
+ "step": 1180
737
+ },
738
+ {
739
+ "epoch": 7.99,
740
+ "learning_rate": 0.0004910706724949438,
741
+ "loss": 2.5858,
742
+ "step": 1190
743
+ },
744
+ {
745
+ "epoch": 8.05,
746
+ "learning_rate": 0.0004906673888040895,
747
+ "loss": 2.5493,
748
+ "step": 1200
749
+ },
750
+ {
751
+ "epoch": 8.12,
752
+ "learning_rate": 0.0004902553712366165,
753
+ "loss": 2.5056,
754
+ "step": 1210
755
+ },
756
+ {
757
+ "epoch": 8.19,
758
+ "learning_rate": 0.0004898346347447405,
759
+ "loss": 2.6127,
760
+ "step": 1220
761
+ },
762
+ {
763
+ "epoch": 8.26,
764
+ "learning_rate": 0.0004894051945970881,
765
+ "loss": 2.519,
766
+ "step": 1230
767
+ },
768
+ {
769
+ "epoch": 8.32,
770
+ "learning_rate": 0.0004889670663781443,
771
+ "loss": 2.4318,
772
+ "step": 1240
773
+ },
774
+ {
775
+ "epoch": 8.39,
776
+ "learning_rate": 0.0004885202659876861,
777
+ "loss": 2.4674,
778
+ "step": 1250
779
+ },
780
+ {
781
+ "epoch": 8.46,
782
+ "learning_rate": 0.00048806480964020554,
783
+ "loss": 2.5532,
784
+ "step": 1260
785
+ },
786
+ {
787
+ "epoch": 8.52,
788
+ "learning_rate": 0.0004876007138643216,
789
+ "loss": 2.5892,
790
+ "step": 1270
791
+ },
792
+ {
793
+ "epoch": 8.59,
794
+ "learning_rate": 0.00048712799550218,
795
+ "loss": 2.5353,
796
+ "step": 1280
797
+ },
798
+ {
799
+ "epoch": 8.66,
800
+ "learning_rate": 0.00048664667170884217,
801
+ "loss": 2.4692,
802
+ "step": 1290
803
+ },
804
+ {
805
+ "epoch": 8.72,
806
+ "learning_rate": 0.0004861567599516628,
807
+ "loss": 2.5329,
808
+ "step": 1300
809
+ },
810
+ {
811
+ "epoch": 8.79,
812
+ "learning_rate": 0.0004856582780096558,
813
+ "loss": 2.5185,
814
+ "step": 1310
815
+ },
816
+ {
817
+ "epoch": 8.86,
818
+ "learning_rate": 0.0004851512439728489,
819
+ "loss": 2.509,
820
+ "step": 1320
821
+ },
822
+ {
823
+ "epoch": 8.93,
824
+ "learning_rate": 0.0004846356762416276,
825
+ "loss": 2.4121,
826
+ "step": 1330
827
+ },
828
+ {
829
+ "epoch": 8.99,
830
+ "learning_rate": 0.00048411159352606734,
831
+ "loss": 2.528,
832
+ "step": 1340
833
+ },
834
+ {
835
+ "epoch": 9.06,
836
+ "learning_rate": 0.0004835790148452539,
837
+ "loss": 2.57,
838
+ "step": 1350
839
+ },
840
+ {
841
+ "epoch": 9.13,
842
+ "learning_rate": 0.000483037959526594,
843
+ "loss": 2.3809,
844
+ "step": 1360
845
+ },
846
+ {
847
+ "epoch": 9.19,
848
+ "learning_rate": 0.0004824884472051136,
849
+ "loss": 2.4399,
850
+ "step": 1370
851
+ },
852
+ {
853
+ "epoch": 9.26,
854
+ "learning_rate": 0.000481930497822745,
855
+ "loss": 2.4777,
856
+ "step": 1380
857
+ },
858
+ {
859
+ "epoch": 9.33,
860
+ "learning_rate": 0.00048136413162760375,
861
+ "loss": 2.6037,
862
+ "step": 1390
863
+ },
864
+ {
865
+ "epoch": 9.4,
866
+ "learning_rate": 0.00048078936917325335,
867
+ "loss": 2.4826,
868
+ "step": 1400
869
+ },
870
+ {
871
+ "epoch": 9.46,
872
+ "learning_rate": 0.0004802062313179595,
873
+ "loss": 2.5202,
874
+ "step": 1410
875
+ },
876
+ {
877
+ "epoch": 9.53,
878
+ "learning_rate": 0.0004796147392239334,
879
+ "loss": 2.5156,
880
+ "step": 1420
881
+ },
882
+ {
883
+ "epoch": 9.6,
884
+ "learning_rate": 0.00047901491435656333,
885
+ "loss": 2.4923,
886
+ "step": 1430
887
+ },
888
+ {
889
+ "epoch": 9.66,
890
+ "learning_rate": 0.000478406778483636,
891
+ "loss": 2.4254,
892
+ "step": 1440
893
+ },
894
+ {
895
+ "epoch": 9.73,
896
+ "learning_rate": 0.0004777903536745466,
897
+ "loss": 2.5741,
898
+ "step": 1450
899
+ },
900
+ {
901
+ "epoch": 9.8,
902
+ "learning_rate": 0.0004771656622994974,
903
+ "loss": 2.5019,
904
+ "step": 1460
905
+ },
906
+ {
907
+ "epoch": 9.87,
908
+ "learning_rate": 0.00047653272702868676,
909
+ "loss": 2.5268,
910
+ "step": 1470
911
+ },
912
+ {
913
+ "epoch": 9.93,
914
+ "learning_rate": 0.0004758915708314858,
915
+ "loss": 2.4928,
916
+ "step": 1480
917
+ },
918
+ {
919
+ "epoch": 10.0,
920
+ "learning_rate": 0.00047524221697560476,
921
+ "loss": 2.5171,
922
+ "step": 1490
923
+ },
924
+ {
925
+ "epoch": 10.07,
926
+ "learning_rate": 0.0004745846890262493,
927
+ "loss": 2.3321,
928
+ "step": 1500
929
+ },
930
+ {
931
+ "epoch": 10.07,
932
+ "eval_accuracy": 0.5985169410705566,
933
+ "eval_accuracy_top5": 0.7327859997749329,
934
+ "eval_loss": 1.7438325881958008,
935
+ "eval_runtime": 103.1085,
936
+ "eval_samples_per_second": 36.69,
937
+ "eval_steps_per_second": 4.587,
938
+ "step": 1500
939
+ },
940
+ {
941
+ "epoch": 10.13,
942
+ "learning_rate": 0.0004739190108452645,
943
+ "loss": 2.4042,
944
+ "step": 1510
945
+ },
946
+ {
947
+ "epoch": 10.2,
948
+ "learning_rate": 0.0004732452065902695,
949
+ "loss": 2.4647,
950
+ "step": 1520
951
+ },
952
+ {
953
+ "epoch": 10.27,
954
+ "learning_rate": 0.0004725633007137805,
955
+ "loss": 2.4076,
956
+ "step": 1530
957
+ },
958
+ {
959
+ "epoch": 10.34,
960
+ "learning_rate": 0.00047187331796232345,
961
+ "loss": 2.2986,
962
+ "step": 1540
963
+ },
964
+ {
965
+ "epoch": 10.4,
966
+ "learning_rate": 0.00047117528337553615,
967
+ "loss": 2.5126,
968
+ "step": 1550
969
+ },
970
+ {
971
+ "epoch": 10.47,
972
+ "learning_rate": 0.00047046922228525926,
973
+ "loss": 2.4857,
974
+ "step": 1560
975
+ },
976
+ {
977
+ "epoch": 10.54,
978
+ "learning_rate": 0.0004697551603146171,
979
+ "loss": 2.4434,
980
+ "step": 1570
981
+ },
982
+ {
983
+ "epoch": 10.6,
984
+ "learning_rate": 0.000469033123377088,
985
+ "loss": 2.3844,
986
+ "step": 1580
987
+ },
988
+ {
989
+ "epoch": 10.67,
990
+ "learning_rate": 0.00046830313767556365,
991
+ "loss": 2.4232,
992
+ "step": 1590
993
+ },
994
+ {
995
+ "epoch": 10.74,
996
+ "learning_rate": 0.0004675652297013984,
997
+ "loss": 2.4638,
998
+ "step": 1600
999
+ },
1000
+ {
1001
+ "epoch": 10.81,
1002
+ "learning_rate": 0.0004668194262334475,
1003
+ "loss": 2.3976,
1004
+ "step": 1610
1005
+ },
1006
+ {
1007
+ "epoch": 10.87,
1008
+ "learning_rate": 0.0004660657543370958,
1009
+ "loss": 2.407,
1010
+ "step": 1620
1011
+ },
1012
+ {
1013
+ "epoch": 10.94,
1014
+ "learning_rate": 0.0004653042413632751,
1015
+ "loss": 2.42,
1016
+ "step": 1630
1017
+ },
1018
+ {
1019
+ "epoch": 11.01,
1020
+ "learning_rate": 0.0004645349149474718,
1021
+ "loss": 2.4782,
1022
+ "step": 1640
1023
+ },
1024
+ {
1025
+ "epoch": 11.07,
1026
+ "learning_rate": 0.00046375780300872405,
1027
+ "loss": 2.3456,
1028
+ "step": 1650
1029
+ },
1030
+ {
1031
+ "epoch": 11.14,
1032
+ "learning_rate": 0.00046297293374860846,
1033
+ "loss": 2.4317,
1034
+ "step": 1660
1035
+ },
1036
+ {
1037
+ "epoch": 11.21,
1038
+ "learning_rate": 0.00046218033565021647,
1039
+ "loss": 2.315,
1040
+ "step": 1670
1041
+ },
1042
+ {
1043
+ "epoch": 11.28,
1044
+ "learning_rate": 0.0004613800374771212,
1045
+ "loss": 2.3308,
1046
+ "step": 1680
1047
+ },
1048
+ {
1049
+ "epoch": 11.34,
1050
+ "learning_rate": 0.000460572068272333,
1051
+ "loss": 2.4006,
1052
+ "step": 1690
1053
+ },
1054
+ {
1055
+ "epoch": 11.41,
1056
+ "learning_rate": 0.0004597564573572461,
1057
+ "loss": 2.4491,
1058
+ "step": 1700
1059
+ },
1060
+ {
1061
+ "epoch": 11.48,
1062
+ "learning_rate": 0.000458933234330574,
1063
+ "loss": 2.4763,
1064
+ "step": 1710
1065
+ },
1066
+ {
1067
+ "epoch": 11.54,
1068
+ "learning_rate": 0.0004581024290672755,
1069
+ "loss": 2.4014,
1070
+ "step": 1720
1071
+ },
1072
+ {
1073
+ "epoch": 11.61,
1074
+ "learning_rate": 0.00045726407171747075,
1075
+ "loss": 2.5596,
1076
+ "step": 1730
1077
+ },
1078
+ {
1079
+ "epoch": 11.68,
1080
+ "learning_rate": 0.0004564181927053468,
1081
+ "loss": 2.4823,
1082
+ "step": 1740
1083
+ },
1084
+ {
1085
+ "epoch": 11.74,
1086
+ "learning_rate": 0.0004555648227280535,
1087
+ "loss": 2.4451,
1088
+ "step": 1750
1089
+ },
1090
+ {
1091
+ "epoch": 11.81,
1092
+ "learning_rate": 0.00045470399275458984,
1093
+ "loss": 2.4504,
1094
+ "step": 1760
1095
+ },
1096
+ {
1097
+ "epoch": 11.88,
1098
+ "learning_rate": 0.00045383573402467967,
1099
+ "loss": 2.3143,
1100
+ "step": 1770
1101
+ },
1102
+ {
1103
+ "epoch": 11.95,
1104
+ "learning_rate": 0.00045296007804763815,
1105
+ "loss": 2.2971,
1106
+ "step": 1780
1107
+ },
1108
+ {
1109
+ "epoch": 12.01,
1110
+ "learning_rate": 0.00045207705660122853,
1111
+ "loss": 2.2387,
1112
+ "step": 1790
1113
+ },
1114
+ {
1115
+ "epoch": 12.08,
1116
+ "learning_rate": 0.00045118670173050846,
1117
+ "loss": 2.386,
1118
+ "step": 1800
1119
+ },
1120
+ {
1121
+ "epoch": 12.15,
1122
+ "learning_rate": 0.0004502890457466673,
1123
+ "loss": 2.3928,
1124
+ "step": 1810
1125
+ },
1126
+ {
1127
+ "epoch": 12.21,
1128
+ "learning_rate": 0.00044938412122585373,
1129
+ "loss": 2.3766,
1130
+ "step": 1820
1131
+ },
1132
+ {
1133
+ "epoch": 12.28,
1134
+ "learning_rate": 0.00044847196100799306,
1135
+ "loss": 2.3229,
1136
+ "step": 1830
1137
+ },
1138
+ {
1139
+ "epoch": 12.35,
1140
+ "learning_rate": 0.000447552598195596,
1141
+ "loss": 2.1737,
1142
+ "step": 1840
1143
+ },
1144
+ {
1145
+ "epoch": 12.42,
1146
+ "learning_rate": 0.00044662606615255705,
1147
+ "loss": 2.3663,
1148
+ "step": 1850
1149
+ },
1150
+ {
1151
+ "epoch": 12.48,
1152
+ "learning_rate": 0.0004456923985029439,
1153
+ "loss": 2.3396,
1154
+ "step": 1860
1155
+ },
1156
+ {
1157
+ "epoch": 12.55,
1158
+ "learning_rate": 0.0004447516291297768,
1159
+ "loss": 2.5236,
1160
+ "step": 1870
1161
+ },
1162
+ {
1163
+ "epoch": 12.62,
1164
+ "learning_rate": 0.0004438037921737997,
1165
+ "loss": 2.242,
1166
+ "step": 1880
1167
+ },
1168
+ {
1169
+ "epoch": 12.68,
1170
+ "learning_rate": 0.00044284892203224024,
1171
+ "loss": 2.3837,
1172
+ "step": 1890
1173
+ },
1174
+ {
1175
+ "epoch": 12.75,
1176
+ "learning_rate": 0.00044188705335756253,
1177
+ "loss": 2.3464,
1178
+ "step": 1900
1179
+ },
1180
+ {
1181
+ "epoch": 12.82,
1182
+ "learning_rate": 0.0004409182210562087,
1183
+ "loss": 2.2084,
1184
+ "step": 1910
1185
+ },
1186
+ {
1187
+ "epoch": 12.89,
1188
+ "learning_rate": 0.00043994246028733296,
1189
+ "loss": 2.3124,
1190
+ "step": 1920
1191
+ },
1192
+ {
1193
+ "epoch": 12.95,
1194
+ "learning_rate": 0.0004389598064615249,
1195
+ "loss": 2.1847,
1196
+ "step": 1930
1197
+ },
1198
+ {
1199
+ "epoch": 13.02,
1200
+ "learning_rate": 0.0004379702952395249,
1201
+ "loss": 2.2371,
1202
+ "step": 1940
1203
+ },
1204
+ {
1205
+ "epoch": 13.09,
1206
+ "learning_rate": 0.00043697396253093,
1207
+ "loss": 2.383,
1208
+ "step": 1950
1209
+ },
1210
+ {
1211
+ "epoch": 13.15,
1212
+ "learning_rate": 0.0004359708444928904,
1213
+ "loss": 2.2143,
1214
+ "step": 1960
1215
+ },
1216
+ {
1217
+ "epoch": 13.22,
1218
+ "learning_rate": 0.0004349609775287977,
1219
+ "loss": 2.2602,
1220
+ "step": 1970
1221
+ },
1222
+ {
1223
+ "epoch": 13.29,
1224
+ "learning_rate": 0.0004339443982869634,
1225
+ "loss": 2.2835,
1226
+ "step": 1980
1227
+ },
1228
+ {
1229
+ "epoch": 13.36,
1230
+ "learning_rate": 0.00043292114365928934,
1231
+ "loss": 2.2528,
1232
+ "step": 1990
1233
+ },
1234
+ {
1235
+ "epoch": 13.42,
1236
+ "learning_rate": 0.00043189125077992867,
1237
+ "loss": 2.3348,
1238
+ "step": 2000
1239
+ },
1240
+ {
1241
+ "epoch": 13.42,
1242
+ "eval_accuracy": 0.6295021176338196,
1243
+ "eval_accuracy_top5": 0.7356991767883301,
1244
+ "eval_loss": 1.62944495677948,
1245
+ "eval_runtime": 99.1269,
1246
+ "eval_samples_per_second": 38.163,
1247
+ "eval_steps_per_second": 4.772,
1248
+ "step": 2000
1249
+ },
1250
+ {
1251
+ "epoch": 13.49,
1252
+ "learning_rate": 0.00043085475702393827,
1253
+ "loss": 2.3693,
1254
+ "step": 2010
1255
+ },
1256
+ {
1257
+ "epoch": 13.56,
1258
+ "learning_rate": 0.0004298117000059223,
1259
+ "loss": 2.2296,
1260
+ "step": 2020
1261
+ },
1262
+ {
1263
+ "epoch": 13.62,
1264
+ "learning_rate": 0.0004287621175786674,
1265
+ "loss": 2.2473,
1266
+ "step": 2030
1267
+ },
1268
+ {
1269
+ "epoch": 13.69,
1270
+ "learning_rate": 0.0004277060478317687,
1271
+ "loss": 2.2507,
1272
+ "step": 2040
1273
+ },
1274
+ {
1275
+ "epoch": 13.76,
1276
+ "learning_rate": 0.00042664352909024784,
1277
+ "loss": 2.3107,
1278
+ "step": 2050
1279
+ },
1280
+ {
1281
+ "epoch": 13.83,
1282
+ "learning_rate": 0.00042557459991316197,
1283
+ "loss": 2.3385,
1284
+ "step": 2060
1285
+ },
1286
+ {
1287
+ "epoch": 13.89,
1288
+ "learning_rate": 0.0004244992990922043,
1289
+ "loss": 2.2102,
1290
+ "step": 2070
1291
+ },
1292
+ {
1293
+ "epoch": 13.96,
1294
+ "learning_rate": 0.0004234176656502968,
1295
+ "loss": 2.3437,
1296
+ "step": 2080
1297
+ },
1298
+ {
1299
+ "epoch": 14.03,
1300
+ "learning_rate": 0.00042232973884017355,
1301
+ "loss": 2.3511,
1302
+ "step": 2090
1303
+ },
1304
+ {
1305
+ "epoch": 14.09,
1306
+ "learning_rate": 0.00042123555814295655,
1307
+ "loss": 2.3496,
1308
+ "step": 2100
1309
+ },
1310
+ {
1311
+ "epoch": 14.16,
1312
+ "learning_rate": 0.00042013516326672276,
1313
+ "loss": 2.2043,
1314
+ "step": 2110
1315
+ },
1316
+ {
1317
+ "epoch": 14.23,
1318
+ "learning_rate": 0.0004190285941450632,
1319
+ "loss": 2.2971,
1320
+ "step": 2120
1321
+ },
1322
+ {
1323
+ "epoch": 14.3,
1324
+ "learning_rate": 0.00041791589093563385,
1325
+ "loss": 2.2116,
1326
+ "step": 2130
1327
+ },
1328
+ {
1329
+ "epoch": 14.36,
1330
+ "learning_rate": 0.000416797094018698,
1331
+ "loss": 2.2991,
1332
+ "step": 2140
1333
+ },
1334
+ {
1335
+ "epoch": 14.43,
1336
+ "learning_rate": 0.00041567224399566104,
1337
+ "loss": 2.2796,
1338
+ "step": 2150
1339
+ },
1340
+ {
1341
+ "epoch": 14.5,
1342
+ "learning_rate": 0.00041454138168759726,
1343
+ "loss": 2.3402,
1344
+ "step": 2160
1345
+ },
1346
+ {
1347
+ "epoch": 14.56,
1348
+ "learning_rate": 0.00041340454813376784,
1349
+ "loss": 2.3073,
1350
+ "step": 2170
1351
+ },
1352
+ {
1353
+ "epoch": 14.63,
1354
+ "learning_rate": 0.0004122617845901322,
1355
+ "loss": 2.2733,
1356
+ "step": 2180
1357
+ },
1358
+ {
1359
+ "epoch": 14.7,
1360
+ "learning_rate": 0.0004111131325278502,
1361
+ "loss": 2.2013,
1362
+ "step": 2190
1363
+ },
1364
+ {
1365
+ "epoch": 14.77,
1366
+ "learning_rate": 0.0004099586336317779,
1367
+ "loss": 2.2275,
1368
+ "step": 2200
1369
+ },
1370
+ {
1371
+ "epoch": 14.83,
1372
+ "learning_rate": 0.00040879832979895376,
1373
+ "loss": 2.2008,
1374
+ "step": 2210
1375
+ },
1376
+ {
1377
+ "epoch": 14.9,
1378
+ "learning_rate": 0.00040763226313707924,
1379
+ "loss": 2.1913,
1380
+ "step": 2220
1381
+ },
1382
+ {
1383
+ "epoch": 14.97,
1384
+ "learning_rate": 0.00040646047596299,
1385
+ "loss": 2.2507,
1386
+ "step": 2230
1387
+ },
1388
+ {
1389
+ "epoch": 15.03,
1390
+ "learning_rate": 0.0004052830108011204,
1391
+ "loss": 2.138,
1392
+ "step": 2240
1393
+ },
1394
+ {
1395
+ "epoch": 15.1,
1396
+ "learning_rate": 0.00040409991038196057,
1397
+ "loss": 2.2183,
1398
+ "step": 2250
1399
+ },
1400
+ {
1401
+ "epoch": 15.17,
1402
+ "learning_rate": 0.00040291121764050533,
1403
+ "loss": 2.2204,
1404
+ "step": 2260
1405
+ },
1406
+ {
1407
+ "epoch": 15.23,
1408
+ "learning_rate": 0.000401716975714696,
1409
+ "loss": 2.1662,
1410
+ "step": 2270
1411
+ },
1412
+ {
1413
+ "epoch": 15.3,
1414
+ "learning_rate": 0.0004005172279438555,
1415
+ "loss": 2.3825,
1416
+ "step": 2280
1417
+ },
1418
+ {
1419
+ "epoch": 15.37,
1420
+ "learning_rate": 0.00039931201786711493,
1421
+ "loss": 2.2073,
1422
+ "step": 2290
1423
+ },
1424
+ {
1425
+ "epoch": 15.44,
1426
+ "learning_rate": 0.0003981013892218336,
1427
+ "loss": 2.1114,
1428
+ "step": 2300
1429
+ },
1430
+ {
1431
+ "epoch": 15.5,
1432
+ "learning_rate": 0.0003968853859420125,
1433
+ "loss": 2.2326,
1434
+ "step": 2310
1435
+ },
1436
+ {
1437
+ "epoch": 15.57,
1438
+ "learning_rate": 0.0003956640521566989,
1439
+ "loss": 2.2042,
1440
+ "step": 2320
1441
+ },
1442
+ {
1443
+ "epoch": 15.64,
1444
+ "learning_rate": 0.00039443743218838553,
1445
+ "loss": 2.1792,
1446
+ "step": 2330
1447
+ },
1448
+ {
1449
+ "epoch": 15.7,
1450
+ "learning_rate": 0.00039320557055140195,
1451
+ "loss": 2.1393,
1452
+ "step": 2340
1453
+ },
1454
+ {
1455
+ "epoch": 15.77,
1456
+ "learning_rate": 0.00039196851195029913,
1457
+ "loss": 2.1661,
1458
+ "step": 2350
1459
+ },
1460
+ {
1461
+ "epoch": 15.84,
1462
+ "learning_rate": 0.00039072630127822707,
1463
+ "loss": 2.2561,
1464
+ "step": 2360
1465
+ },
1466
+ {
1467
+ "epoch": 15.91,
1468
+ "learning_rate": 0.0003894789836153058,
1469
+ "loss": 2.1827,
1470
+ "step": 2370
1471
+ },
1472
+ {
1473
+ "epoch": 15.97,
1474
+ "learning_rate": 0.00038822660422698895,
1475
+ "loss": 2.2583,
1476
+ "step": 2380
1477
+ },
1478
+ {
1479
+ "epoch": 16.04,
1480
+ "learning_rate": 0.00038696920856242176,
1481
+ "loss": 2.1149,
1482
+ "step": 2390
1483
+ },
1484
+ {
1485
+ "epoch": 16.11,
1486
+ "learning_rate": 0.0003857068422527908,
1487
+ "loss": 2.1045,
1488
+ "step": 2400
1489
+ },
1490
+ {
1491
+ "epoch": 16.17,
1492
+ "learning_rate": 0.0003844395511096692,
1493
+ "loss": 2.2076,
1494
+ "step": 2410
1495
+ },
1496
+ {
1497
+ "epoch": 16.24,
1498
+ "learning_rate": 0.00038316738112335294,
1499
+ "loss": 2.2332,
1500
+ "step": 2420
1501
+ },
1502
+ {
1503
+ "epoch": 16.31,
1504
+ "learning_rate": 0.00038189037846119247,
1505
+ "loss": 2.1815,
1506
+ "step": 2430
1507
+ },
1508
+ {
1509
+ "epoch": 16.38,
1510
+ "learning_rate": 0.0003806085894659174,
1511
+ "loss": 2.139,
1512
+ "step": 2440
1513
+ },
1514
+ {
1515
+ "epoch": 16.44,
1516
+ "learning_rate": 0.00037932206065395427,
1517
+ "loss": 2.0076,
1518
+ "step": 2450
1519
+ },
1520
+ {
1521
+ "epoch": 16.51,
1522
+ "learning_rate": 0.0003780308387137387,
1523
+ "loss": 2.2418,
1524
+ "step": 2460
1525
+ },
1526
+ {
1527
+ "epoch": 16.58,
1528
+ "learning_rate": 0.00037673497050402143,
1529
+ "loss": 2.1668,
1530
+ "step": 2470
1531
+ },
1532
+ {
1533
+ "epoch": 16.64,
1534
+ "learning_rate": 0.0003754345030521669,
1535
+ "loss": 2.1749,
1536
+ "step": 2480
1537
+ },
1538
+ {
1539
+ "epoch": 16.71,
1540
+ "learning_rate": 0.0003741294835524475,
1541
+ "loss": 2.1023,
1542
+ "step": 2490
1543
+ },
1544
+ {
1545
+ "epoch": 16.78,
1546
+ "learning_rate": 0.00037281995936433026,
1547
+ "loss": 2.0315,
1548
+ "step": 2500
1549
+ },
1550
+ {
1551
+ "epoch": 16.78,
1552
+ "eval_accuracy": 0.6353284120559692,
1553
+ "eval_accuracy_top5": 0.7513241767883301,
1554
+ "eval_loss": 1.620951771736145,
1555
+ "eval_runtime": 106.1983,
1556
+ "eval_samples_per_second": 35.622,
1557
+ "eval_steps_per_second": 4.454,
1558
+ "step": 2500
1559
+ },
1560
+ {
1561
+ "epoch": 16.85,
1562
+ "learning_rate": 0.0003715059780107586,
1563
+ "loss": 2.1777,
1564
+ "step": 2510
1565
+ },
1566
+ {
1567
+ "epoch": 16.91,
1568
+ "learning_rate": 0.0003701875871764275,
1569
+ "loss": 2.0452,
1570
+ "step": 2520
1571
+ },
1572
+ {
1573
+ "epoch": 16.98,
1574
+ "learning_rate": 0.0003688648347060529,
1575
+ "loss": 2.2884,
1576
+ "step": 2530
1577
+ },
1578
+ {
1579
+ "epoch": 17.05,
1580
+ "learning_rate": 0.00036753776860263594,
1581
+ "loss": 2.1402,
1582
+ "step": 2540
1583
+ },
1584
+ {
1585
+ "epoch": 17.11,
1586
+ "learning_rate": 0.00036620643702571997,
1587
+ "loss": 2.0759,
1588
+ "step": 2550
1589
+ },
1590
+ {
1591
+ "epoch": 17.18,
1592
+ "learning_rate": 0.00036487088828964377,
1593
+ "loss": 2.0357,
1594
+ "step": 2560
1595
+ },
1596
+ {
1597
+ "epoch": 17.25,
1598
+ "learning_rate": 0.00036353117086178777,
1599
+ "loss": 2.0975,
1600
+ "step": 2570
1601
+ },
1602
+ {
1603
+ "epoch": 17.32,
1604
+ "learning_rate": 0.00036218733336081506,
1605
+ "loss": 2.2448,
1606
+ "step": 2580
1607
+ },
1608
+ {
1609
+ "epoch": 17.38,
1610
+ "learning_rate": 0.00036083942455490735,
1611
+ "loss": 2.1251,
1612
+ "step": 2590
1613
+ },
1614
+ {
1615
+ "epoch": 17.45,
1616
+ "learning_rate": 0.0003594874933599949,
1617
+ "loss": 2.1602,
1618
+ "step": 2600
1619
+ },
1620
+ {
1621
+ "epoch": 17.52,
1622
+ "learning_rate": 0.00035813158883798134,
1623
+ "loss": 2.1744,
1624
+ "step": 2610
1625
+ },
1626
+ {
1627
+ "epoch": 17.58,
1628
+ "learning_rate": 0.00035677176019496343,
1629
+ "loss": 2.1631,
1630
+ "step": 2620
1631
+ },
1632
+ {
1633
+ "epoch": 17.65,
1634
+ "learning_rate": 0.0003554080567794451,
1635
+ "loss": 2.1669,
1636
+ "step": 2630
1637
+ },
1638
+ {
1639
+ "epoch": 17.72,
1640
+ "learning_rate": 0.0003540405280805467,
1641
+ "loss": 2.1181,
1642
+ "step": 2640
1643
+ },
1644
+ {
1645
+ "epoch": 17.79,
1646
+ "learning_rate": 0.00035266922372620906,
1647
+ "loss": 2.0952,
1648
+ "step": 2650
1649
+ },
1650
+ {
1651
+ "epoch": 17.85,
1652
+ "learning_rate": 0.00035129419348139246,
1653
+ "loss": 2.1549,
1654
+ "step": 2660
1655
+ },
1656
+ {
1657
+ "epoch": 17.92,
1658
+ "learning_rate": 0.0003499154872462705,
1659
+ "loss": 2.1738,
1660
+ "step": 2670
1661
+ },
1662
+ {
1663
+ "epoch": 17.99,
1664
+ "learning_rate": 0.0003485331550544197,
1665
+ "loss": 2.1793,
1666
+ "step": 2680
1667
+ },
1668
+ {
1669
+ "epoch": 18.05,
1670
+ "learning_rate": 0.00034714724707100283,
1671
+ "loss": 2.1329,
1672
+ "step": 2690
1673
+ },
1674
+ {
1675
+ "epoch": 18.12,
1676
+ "learning_rate": 0.00034575781359094947,
1677
+ "loss": 2.0711,
1678
+ "step": 2700
1679
+ },
1680
+ {
1681
+ "epoch": 18.19,
1682
+ "learning_rate": 0.0003443649050371301,
1683
+ "loss": 2.0466,
1684
+ "step": 2710
1685
+ },
1686
+ {
1687
+ "epoch": 18.26,
1688
+ "learning_rate": 0.00034296857195852647,
1689
+ "loss": 2.1244,
1690
+ "step": 2720
1691
+ },
1692
+ {
1693
+ "epoch": 18.32,
1694
+ "learning_rate": 0.0003415688650283973,
1695
+ "loss": 2.1518,
1696
+ "step": 2730
1697
+ },
1698
+ {
1699
+ "epoch": 18.39,
1700
+ "learning_rate": 0.0003401658350424389,
1701
+ "loss": 2.1096,
1702
+ "step": 2740
1703
+ },
1704
+ {
1705
+ "epoch": 18.46,
1706
+ "learning_rate": 0.0003387595329169423,
1707
+ "loss": 2.0938,
1708
+ "step": 2750
1709
+ },
1710
+ {
1711
+ "epoch": 18.52,
1712
+ "learning_rate": 0.0003373500096869451,
1713
+ "loss": 2.1085,
1714
+ "step": 2760
1715
+ },
1716
+ {
1717
+ "epoch": 18.59,
1718
+ "learning_rate": 0.0003359373165043796,
1719
+ "loss": 2.0453,
1720
+ "step": 2770
1721
+ },
1722
+ {
1723
+ "epoch": 18.66,
1724
+ "learning_rate": 0.0003345215046362165,
1725
+ "loss": 2.2149,
1726
+ "step": 2780
1727
+ },
1728
+ {
1729
+ "epoch": 18.72,
1730
+ "learning_rate": 0.00033310262546260434,
1731
+ "loss": 2.1468,
1732
+ "step": 2790
1733
+ },
1734
+ {
1735
+ "epoch": 18.79,
1736
+ "learning_rate": 0.0003316807304750047,
1737
+ "loss": 2.0384,
1738
+ "step": 2800
1739
+ },
1740
+ {
1741
+ "epoch": 18.86,
1742
+ "learning_rate": 0.0003302558712743241,
1743
+ "loss": 2.0865,
1744
+ "step": 2810
1745
+ },
1746
+ {
1747
+ "epoch": 18.93,
1748
+ "learning_rate": 0.00032882809956904065,
1749
+ "loss": 2.1671,
1750
+ "step": 2820
1751
+ },
1752
+ {
1753
+ "epoch": 18.99,
1754
+ "learning_rate": 0.0003273974671733283,
1755
+ "loss": 2.0147,
1756
+ "step": 2830
1757
+ },
1758
+ {
1759
+ "epoch": 19.06,
1760
+ "learning_rate": 0.0003259640260051759,
1761
+ "loss": 1.9943,
1762
+ "step": 2840
1763
+ },
1764
+ {
1765
+ "epoch": 19.13,
1766
+ "learning_rate": 0.00032452782808450355,
1767
+ "loss": 1.9631,
1768
+ "step": 2850
1769
+ },
1770
+ {
1771
+ "epoch": 19.19,
1772
+ "learning_rate": 0.00032308892553127425,
1773
+ "loss": 2.1063,
1774
+ "step": 2860
1775
+ },
1776
+ {
1777
+ "epoch": 19.26,
1778
+ "learning_rate": 0.00032164737056360307,
1779
+ "loss": 1.9979,
1780
+ "step": 2870
1781
+ },
1782
+ {
1783
+ "epoch": 19.33,
1784
+ "learning_rate": 0.0003202032154958615,
1785
+ "loss": 1.9259,
1786
+ "step": 2880
1787
+ },
1788
+ {
1789
+ "epoch": 19.4,
1790
+ "learning_rate": 0.0003187565127367797,
1791
+ "loss": 1.9233,
1792
+ "step": 2890
1793
+ },
1794
+ {
1795
+ "epoch": 19.46,
1796
+ "learning_rate": 0.00031730731478754365,
1797
+ "loss": 2.0203,
1798
+ "step": 2900
1799
+ },
1800
+ {
1801
+ "epoch": 19.53,
1802
+ "learning_rate": 0.00031585567423989084,
1803
+ "loss": 2.0238,
1804
+ "step": 2910
1805
+ },
1806
+ {
1807
+ "epoch": 19.6,
1808
+ "learning_rate": 0.00031440164377420085,
1809
+ "loss": 2.0497,
1810
+ "step": 2920
1811
+ },
1812
+ {
1813
+ "epoch": 19.66,
1814
+ "learning_rate": 0.0003129452761575843,
1815
+ "loss": 1.9399,
1816
+ "step": 2930
1817
+ },
1818
+ {
1819
+ "epoch": 19.73,
1820
+ "learning_rate": 0.00031148662424196723,
1821
+ "loss": 2.0291,
1822
+ "step": 2940
1823
+ },
1824
+ {
1825
+ "epoch": 19.8,
1826
+ "learning_rate": 0.0003100257409621738,
1827
+ "loss": 2.0845,
1828
+ "step": 2950
1829
+ },
1830
+ {
1831
+ "epoch": 19.87,
1832
+ "learning_rate": 0.0003085626793340045,
1833
+ "loss": 1.9692,
1834
+ "step": 2960
1835
+ },
1836
+ {
1837
+ "epoch": 19.93,
1838
+ "learning_rate": 0.00030709749245231297,
1839
+ "loss": 2.0479,
1840
+ "step": 2970
1841
+ },
1842
+ {
1843
+ "epoch": 20.0,
1844
+ "learning_rate": 0.0003056302334890786,
1845
+ "loss": 2.0145,
1846
+ "step": 2980
1847
+ },
1848
+ {
1849
+ "epoch": 20.07,
1850
+ "learning_rate": 0.00030416095569147717,
1851
+ "loss": 1.9802,
1852
+ "step": 2990
1853
+ },
1854
+ {
1855
+ "epoch": 20.13,
1856
+ "learning_rate": 0.00030268971237994835,
1857
+ "loss": 2.0705,
1858
+ "step": 3000
1859
+ },
1860
+ {
1861
+ "epoch": 20.13,
1862
+ "eval_accuracy": 0.6464512944221497,
1863
+ "eval_accuracy_top5": 0.7627118825912476,
1864
+ "eval_loss": 1.5695319175720215,
1865
+ "eval_runtime": 101.6469,
1866
+ "eval_samples_per_second": 37.217,
1867
+ "eval_steps_per_second": 4.653,
1868
+ "step": 3000
1869
+ },
1870
+ {
1871
+ "epoch": 20.2,
1872
+ "learning_rate": 0.00030121655694626096,
1873
+ "loss": 1.9999,
1874
+ "step": 3010
1875
+ },
1876
+ {
1877
+ "epoch": 20.27,
1878
+ "learning_rate": 0.00029974154285157496,
1879
+ "loss": 2.0053,
1880
+ "step": 3020
1881
+ },
1882
+ {
1883
+ "epoch": 20.34,
1884
+ "learning_rate": 0.0002982647236245017,
1885
+ "loss": 1.9484,
1886
+ "step": 3030
1887
+ },
1888
+ {
1889
+ "epoch": 20.4,
1890
+ "learning_rate": 0.00029678615285916126,
1891
+ "loss": 1.9829,
1892
+ "step": 3040
1893
+ },
1894
+ {
1895
+ "epoch": 20.47,
1896
+ "learning_rate": 0.0002953058842132373,
1897
+ "loss": 1.9263,
1898
+ "step": 3050
1899
+ },
1900
+ {
1901
+ "epoch": 20.54,
1902
+ "learning_rate": 0.00029382397140603013,
1903
+ "loss": 1.955,
1904
+ "step": 3060
1905
+ },
1906
+ {
1907
+ "epoch": 20.6,
1908
+ "learning_rate": 0.00029234046821650694,
1909
+ "loss": 2.0098,
1910
+ "step": 3070
1911
+ },
1912
+ {
1913
+ "epoch": 20.67,
1914
+ "learning_rate": 0.0002908554284813503,
1915
+ "loss": 1.9278,
1916
+ "step": 3080
1917
+ },
1918
+ {
1919
+ "epoch": 20.74,
1920
+ "learning_rate": 0.00028936890609300447,
1921
+ "loss": 2.0939,
1922
+ "step": 3090
1923
+ },
1924
+ {
1925
+ "epoch": 20.81,
1926
+ "learning_rate": 0.00028788095499771943,
1927
+ "loss": 1.9715,
1928
+ "step": 3100
1929
+ },
1930
+ {
1931
+ "epoch": 20.87,
1932
+ "learning_rate": 0.0002863916291935933,
1933
+ "loss": 2.0042,
1934
+ "step": 3110
1935
+ },
1936
+ {
1937
+ "epoch": 20.94,
1938
+ "learning_rate": 0.00028490098272861283,
1939
+ "loss": 2.0696,
1940
+ "step": 3120
1941
+ },
1942
+ {
1943
+ "epoch": 21.01,
1944
+ "learning_rate": 0.0002834090696986916,
1945
+ "loss": 2.0059,
1946
+ "step": 3130
1947
+ },
1948
+ {
1949
+ "epoch": 21.07,
1950
+ "learning_rate": 0.00028191594424570754,
1951
+ "loss": 1.9319,
1952
+ "step": 3140
1953
+ },
1954
+ {
1955
+ "epoch": 21.14,
1956
+ "learning_rate": 0.00028042166055553723,
1957
+ "loss": 2.0266,
1958
+ "step": 3150
1959
+ },
1960
+ {
1961
+ "epoch": 21.21,
1962
+ "learning_rate": 0.00027892627285609035,
1963
+ "loss": 1.9712,
1964
+ "step": 3160
1965
+ },
1966
+ {
1967
+ "epoch": 21.28,
1968
+ "learning_rate": 0.0002774298354153411,
1969
+ "loss": 1.9637,
1970
+ "step": 3170
1971
+ },
1972
+ {
1973
+ "epoch": 21.34,
1974
+ "learning_rate": 0.0002759324025393591,
1975
+ "loss": 1.9696,
1976
+ "step": 3180
1977
+ },
1978
+ {
1979
+ "epoch": 21.41,
1980
+ "learning_rate": 0.00027443402857033864,
1981
+ "loss": 1.9457,
1982
+ "step": 3190
1983
+ },
1984
+ {
1985
+ "epoch": 21.48,
1986
+ "learning_rate": 0.00027293476788462623,
1987
+ "loss": 1.9282,
1988
+ "step": 3200
1989
+ },
1990
+ {
1991
+ "epoch": 21.54,
1992
+ "learning_rate": 0.000271434674890748,
1993
+ "loss": 2.0554,
1994
+ "step": 3210
1995
+ },
1996
+ {
1997
+ "epoch": 21.61,
1998
+ "learning_rate": 0.0002699338040274343,
1999
+ "loss": 1.9161,
2000
+ "step": 3220
2001
+ },
2002
+ {
2003
+ "epoch": 21.68,
2004
+ "learning_rate": 0.0002684322097616448,
2005
+ "loss": 1.9126,
2006
+ "step": 3230
2007
+ },
2008
+ {
2009
+ "epoch": 21.74,
2010
+ "learning_rate": 0.0002669299465865914,
2011
+ "loss": 2.0167,
2012
+ "step": 3240
2013
+ },
2014
+ {
2015
+ "epoch": 21.81,
2016
+ "learning_rate": 0.00026542706901976125,
2017
+ "loss": 1.9754,
2018
+ "step": 3250
2019
+ },
2020
+ {
2021
+ "epoch": 21.88,
2022
+ "learning_rate": 0.00026392363160093746,
2023
+ "loss": 1.9503,
2024
+ "step": 3260
2025
+ },
2026
+ {
2027
+ "epoch": 21.95,
2028
+ "learning_rate": 0.00026241968889022065,
2029
+ "loss": 2.0577,
2030
+ "step": 3270
2031
+ },
2032
+ {
2033
+ "epoch": 22.01,
2034
+ "learning_rate": 0.0002609152954660484,
2035
+ "loss": 1.9624,
2036
+ "step": 3280
2037
+ },
2038
+ {
2039
+ "epoch": 22.08,
2040
+ "learning_rate": 0.000259410505923215,
2041
+ "loss": 1.9243,
2042
+ "step": 3290
2043
+ },
2044
+ {
2045
+ "epoch": 22.15,
2046
+ "learning_rate": 0.0002579053748708897,
2047
+ "loss": 1.9068,
2048
+ "step": 3300
2049
+ },
2050
+ {
2051
+ "epoch": 22.21,
2052
+ "learning_rate": 0.0002563999569306355,
2053
+ "loss": 1.8909,
2054
+ "step": 3310
2055
+ },
2056
+ {
2057
+ "epoch": 22.28,
2058
+ "learning_rate": 0.00025489430673442635,
2059
+ "loss": 1.9244,
2060
+ "step": 3320
2061
+ },
2062
+ {
2063
+ "epoch": 22.35,
2064
+ "learning_rate": 0.0002533884789226651,
2065
+ "loss": 1.9617,
2066
+ "step": 3330
2067
+ },
2068
+ {
2069
+ "epoch": 22.42,
2070
+ "learning_rate": 0.00025188252814220004,
2071
+ "loss": 1.8954,
2072
+ "step": 3340
2073
+ },
2074
+ {
2075
+ "epoch": 22.48,
2076
+ "learning_rate": 0.0002503765090443422,
2077
+ "loss": 1.95,
2078
+ "step": 3350
2079
+ },
2080
+ {
2081
+ "epoch": 22.55,
2082
+ "learning_rate": 0.0002488704762828819,
2083
+ "loss": 1.9047,
2084
+ "step": 3360
2085
+ },
2086
+ {
2087
+ "epoch": 22.62,
2088
+ "learning_rate": 0.0002473644845121051,
2089
+ "loss": 2.0102,
2090
+ "step": 3370
2091
+ },
2092
+ {
2093
+ "epoch": 22.68,
2094
+ "learning_rate": 0.00024585858838481055,
2095
+ "loss": 1.9811,
2096
+ "step": 3380
2097
+ },
2098
+ {
2099
+ "epoch": 22.75,
2100
+ "learning_rate": 0.0002443528425503256,
2101
+ "loss": 1.8512,
2102
+ "step": 3390
2103
+ },
2104
+ {
2105
+ "epoch": 22.82,
2106
+ "learning_rate": 0.00024284730165252387,
2107
+ "loss": 1.929,
2108
+ "step": 3400
2109
+ },
2110
+ {
2111
+ "epoch": 22.89,
2112
+ "learning_rate": 0.00024134202032784156,
2113
+ "loss": 1.8707,
2114
+ "step": 3410
2115
+ },
2116
+ {
2117
+ "epoch": 22.95,
2118
+ "learning_rate": 0.00023983705320329508,
2119
+ "loss": 1.8647,
2120
+ "step": 3420
2121
+ },
2122
+ {
2123
+ "epoch": 23.02,
2124
+ "learning_rate": 0.00023833245489449805,
2125
+ "loss": 1.9206,
2126
+ "step": 3430
2127
+ },
2128
+ {
2129
+ "epoch": 23.09,
2130
+ "learning_rate": 0.00023682828000368012,
2131
+ "loss": 1.8319,
2132
+ "step": 3440
2133
+ },
2134
+ {
2135
+ "epoch": 23.15,
2136
+ "learning_rate": 0.00023532458311770486,
2137
+ "loss": 1.8411,
2138
+ "step": 3450
2139
+ },
2140
+ {
2141
+ "epoch": 23.22,
2142
+ "learning_rate": 0.0002338214188060887,
2143
+ "loss": 1.7666,
2144
+ "step": 3460
2145
+ },
2146
+ {
2147
+ "epoch": 23.29,
2148
+ "learning_rate": 0.0002323188416190211,
2149
+ "loss": 1.8131,
2150
+ "step": 3470
2151
+ },
2152
+ {
2153
+ "epoch": 23.36,
2154
+ "learning_rate": 0.00023081690608538463,
2155
+ "loss": 1.9416,
2156
+ "step": 3480
2157
+ },
2158
+ {
2159
+ "epoch": 23.42,
2160
+ "learning_rate": 0.00022931566671077603,
2161
+ "loss": 1.8075,
2162
+ "step": 3490
2163
+ },
2164
+ {
2165
+ "epoch": 23.49,
2166
+ "learning_rate": 0.0002278151779755281,
2167
+ "loss": 1.9087,
2168
+ "step": 3500
2169
+ },
2170
+ {
2171
+ "epoch": 23.49,
2172
+ "eval_accuracy": 0.6922669410705566,
2173
+ "eval_accuracy_top5": 0.7862817645072937,
2174
+ "eval_loss": 1.3863152265548706,
2175
+ "eval_runtime": 99.8777,
2176
+ "eval_samples_per_second": 37.876,
2177
+ "eval_steps_per_second": 4.736,
2178
+ "step": 3500
2179
+ },
2180
+ {
2181
+ "epoch": 23.56,
2182
+ "learning_rate": 0.00022631549433273294,
2183
+ "loss": 1.9428,
2184
+ "step": 3510
2185
+ },
2186
+ {
2187
+ "epoch": 23.62,
2188
+ "learning_rate": 0.0002248166702062657,
2189
+ "loss": 1.8847,
2190
+ "step": 3520
2191
+ },
2192
+ {
2193
+ "epoch": 23.69,
2194
+ "learning_rate": 0.00022331875998880943,
2195
+ "loss": 1.9183,
2196
+ "step": 3530
2197
+ },
2198
+ {
2199
+ "epoch": 23.76,
2200
+ "learning_rate": 0.0002218218180398811,
2201
+ "loss": 1.9658,
2202
+ "step": 3540
2203
+ },
2204
+ {
2205
+ "epoch": 23.83,
2206
+ "learning_rate": 0.0002203258986838591,
2207
+ "loss": 1.8596,
2208
+ "step": 3550
2209
+ },
2210
+ {
2211
+ "epoch": 23.89,
2212
+ "learning_rate": 0.00021883105620801164,
2213
+ "loss": 1.9158,
2214
+ "step": 3560
2215
+ },
2216
+ {
2217
+ "epoch": 23.96,
2218
+ "learning_rate": 0.0002173373448605267,
2219
+ "loss": 1.8477,
2220
+ "step": 3570
2221
+ },
2222
+ {
2223
+ "epoch": 24.03,
2224
+ "learning_rate": 0.0002158448188485433,
2225
+ "loss": 1.9931,
2226
+ "step": 3580
2227
+ },
2228
+ {
2229
+ "epoch": 24.09,
2230
+ "learning_rate": 0.00021435353233618446,
2231
+ "loss": 1.8643,
2232
+ "step": 3590
2233
+ },
2234
+ {
2235
+ "epoch": 24.16,
2236
+ "learning_rate": 0.00021286353944259138,
2237
+ "loss": 1.8867,
2238
+ "step": 3600
2239
+ },
2240
+ {
2241
+ "epoch": 24.23,
2242
+ "learning_rate": 0.00021137489423995963,
2243
+ "loss": 1.9012,
2244
+ "step": 3610
2245
+ },
2246
+ {
2247
+ "epoch": 24.3,
2248
+ "learning_rate": 0.00020988765075157648,
2249
+ "loss": 1.8438,
2250
+ "step": 3620
2251
+ },
2252
+ {
2253
+ "epoch": 24.36,
2254
+ "learning_rate": 0.00020840186294986097,
2255
+ "loss": 1.8712,
2256
+ "step": 3630
2257
+ },
2258
+ {
2259
+ "epoch": 24.43,
2260
+ "learning_rate": 0.00020691758475440482,
2261
+ "loss": 1.8504,
2262
+ "step": 3640
2263
+ },
2264
+ {
2265
+ "epoch": 24.5,
2266
+ "learning_rate": 0.0002054348700300158,
2267
+ "loss": 1.9375,
2268
+ "step": 3650
2269
+ },
2270
+ {
2271
+ "epoch": 24.56,
2272
+ "learning_rate": 0.00020395377258476278,
2273
+ "loss": 1.8488,
2274
+ "step": 3660
2275
+ },
2276
+ {
2277
+ "epoch": 24.63,
2278
+ "learning_rate": 0.0002024743461680234,
2279
+ "loss": 1.8815,
2280
+ "step": 3670
2281
+ },
2282
+ {
2283
+ "epoch": 24.7,
2284
+ "learning_rate": 0.00020099664446853328,
2285
+ "loss": 1.8098,
2286
+ "step": 3680
2287
+ },
2288
+ {
2289
+ "epoch": 24.77,
2290
+ "learning_rate": 0.00019952072111243738,
2291
+ "loss": 1.8489,
2292
+ "step": 3690
2293
+ },
2294
+ {
2295
+ "epoch": 24.83,
2296
+ "learning_rate": 0.00019804662966134442,
2297
+ "loss": 1.9232,
2298
+ "step": 3700
2299
+ },
2300
+ {
2301
+ "epoch": 24.9,
2302
+ "learning_rate": 0.0001965744236103828,
2303
+ "loss": 1.8177,
2304
+ "step": 3710
2305
+ },
2306
+ {
2307
+ "epoch": 24.97,
2308
+ "learning_rate": 0.0001951041563862593,
2309
+ "loss": 1.9149,
2310
+ "step": 3720
2311
+ },
2312
+ {
2313
+ "epoch": 25.03,
2314
+ "learning_rate": 0.00019363588134532007,
2315
+ "loss": 1.8291,
2316
+ "step": 3730
2317
+ },
2318
+ {
2319
+ "epoch": 25.1,
2320
+ "learning_rate": 0.0001921696517716147,
2321
+ "loss": 1.883,
2322
+ "step": 3740
2323
+ },
2324
+ {
2325
+ "epoch": 25.17,
2326
+ "learning_rate": 0.00019070552087496203,
2327
+ "loss": 1.8461,
2328
+ "step": 3750
2329
+ },
2330
+ {
2331
+ "epoch": 25.23,
2332
+ "learning_rate": 0.0001892435417890197,
2333
+ "loss": 1.8295,
2334
+ "step": 3760
2335
+ },
2336
+ {
2337
+ "epoch": 25.3,
2338
+ "learning_rate": 0.00018778376756935534,
2339
+ "loss": 1.887,
2340
+ "step": 3770
2341
+ },
2342
+ {
2343
+ "epoch": 25.37,
2344
+ "learning_rate": 0.00018632625119152163,
2345
+ "loss": 1.8155,
2346
+ "step": 3780
2347
+ },
2348
+ {
2349
+ "epoch": 25.44,
2350
+ "learning_rate": 0.0001848710455491336,
2351
+ "loss": 1.8593,
2352
+ "step": 3790
2353
+ },
2354
+ {
2355
+ "epoch": 25.5,
2356
+ "learning_rate": 0.00018341820345194932,
2357
+ "loss": 1.7904,
2358
+ "step": 3800
2359
+ },
2360
+ {
2361
+ "epoch": 25.57,
2362
+ "learning_rate": 0.00018196777762395275,
2363
+ "loss": 1.9035,
2364
+ "step": 3810
2365
+ },
2366
+ {
2367
+ "epoch": 25.64,
2368
+ "learning_rate": 0.00018051982070144135,
2369
+ "loss": 1.8704,
2370
+ "step": 3820
2371
+ },
2372
+ {
2373
+ "epoch": 25.7,
2374
+ "learning_rate": 0.00017907438523111528,
2375
+ "loss": 1.7725,
2376
+ "step": 3830
2377
+ },
2378
+ {
2379
+ "epoch": 25.77,
2380
+ "learning_rate": 0.0001776315236681706,
2381
+ "loss": 1.7824,
2382
+ "step": 3840
2383
+ },
2384
+ {
2385
+ "epoch": 25.84,
2386
+ "learning_rate": 0.00017619128837439546,
2387
+ "loss": 1.9004,
2388
+ "step": 3850
2389
+ },
2390
+ {
2391
+ "epoch": 25.91,
2392
+ "learning_rate": 0.00017475373161627035,
2393
+ "loss": 1.7393,
2394
+ "step": 3860
2395
+ },
2396
+ {
2397
+ "epoch": 25.97,
2398
+ "learning_rate": 0.00017331890556307095,
2399
+ "loss": 1.8055,
2400
+ "step": 3870
2401
+ },
2402
+ {
2403
+ "epoch": 26.04,
2404
+ "learning_rate": 0.0001718868622849752,
2405
+ "loss": 1.6775,
2406
+ "step": 3880
2407
+ },
2408
+ {
2409
+ "epoch": 26.11,
2410
+ "learning_rate": 0.00017045765375117299,
2411
+ "loss": 1.765,
2412
+ "step": 3890
2413
+ },
2414
+ {
2415
+ "epoch": 26.17,
2416
+ "learning_rate": 0.00016903133182798115,
2417
+ "loss": 1.7172,
2418
+ "step": 3900
2419
+ },
2420
+ {
2421
+ "epoch": 26.24,
2422
+ "learning_rate": 0.00016760794827696055,
2423
+ "loss": 1.7846,
2424
+ "step": 3910
2425
+ },
2426
+ {
2427
+ "epoch": 26.31,
2428
+ "learning_rate": 0.0001661875547530378,
2429
+ "loss": 1.8185,
2430
+ "step": 3920
2431
+ },
2432
+ {
2433
+ "epoch": 26.38,
2434
+ "learning_rate": 0.0001647702028026308,
2435
+ "loss": 1.7271,
2436
+ "step": 3930
2437
+ },
2438
+ {
2439
+ "epoch": 26.44,
2440
+ "learning_rate": 0.00016335594386177794,
2441
+ "loss": 1.7887,
2442
+ "step": 3940
2443
+ },
2444
+ {
2445
+ "epoch": 26.51,
2446
+ "learning_rate": 0.0001619448292542716,
2447
+ "loss": 1.7533,
2448
+ "step": 3950
2449
+ },
2450
+ {
2451
+ "epoch": 26.58,
2452
+ "learning_rate": 0.0001605369101897956,
2453
+ "loss": 1.8131,
2454
+ "step": 3960
2455
+ },
2456
+ {
2457
+ "epoch": 26.64,
2458
+ "learning_rate": 0.00015913223776206666,
2459
+ "loss": 1.7736,
2460
+ "step": 3970
2461
+ },
2462
+ {
2463
+ "epoch": 26.71,
2464
+ "learning_rate": 0.00015773086294698036,
2465
+ "loss": 1.7481,
2466
+ "step": 3980
2467
+ },
2468
+ {
2469
+ "epoch": 26.78,
2470
+ "learning_rate": 0.0001563328366007612,
2471
+ "loss": 1.8161,
2472
+ "step": 3990
2473
+ },
2474
+ {
2475
+ "epoch": 26.85,
2476
+ "learning_rate": 0.0001549382094581166,
2477
+ "loss": 1.816,
2478
+ "step": 4000
2479
+ },
2480
+ {
2481
+ "epoch": 26.85,
2482
+ "eval_accuracy": 0.7110699415206909,
2483
+ "eval_accuracy_top5": 0.7913135886192322,
2484
+ "eval_loss": 1.3816050291061401,
2485
+ "eval_runtime": 1281.5963,
2486
+ "eval_samples_per_second": 2.952,
2487
+ "eval_steps_per_second": 0.369,
2488
+ "step": 4000
2489
+ },
2490
+ {
2491
+ "epoch": 26.91,
2492
+ "learning_rate": 0.00015354703213039651,
2493
+ "loss": 1.7483,
2494
+ "step": 4010
2495
+ },
2496
+ {
2497
+ "epoch": 26.98,
2498
+ "learning_rate": 0.00015215935510375607,
2499
+ "loss": 1.7502,
2500
+ "step": 4020
2501
+ },
2502
+ {
2503
+ "epoch": 27.05,
2504
+ "learning_rate": 0.00015077522873732375,
2505
+ "loss": 1.7954,
2506
+ "step": 4030
2507
+ },
2508
+ {
2509
+ "epoch": 27.11,
2510
+ "learning_rate": 0.0001493947032613735,
2511
+ "loss": 1.7957,
2512
+ "step": 4040
2513
+ },
2514
+ {
2515
+ "epoch": 27.18,
2516
+ "learning_rate": 0.00014801782877550234,
2517
+ "loss": 1.7301,
2518
+ "step": 4050
2519
+ },
2520
+ {
2521
+ "epoch": 27.25,
2522
+ "learning_rate": 0.00014664465524681197,
2523
+ "loss": 1.7393,
2524
+ "step": 4060
2525
+ },
2526
+ {
2527
+ "epoch": 27.32,
2528
+ "learning_rate": 0.00014527523250809543,
2529
+ "loss": 1.7667,
2530
+ "step": 4070
2531
+ },
2532
+ {
2533
+ "epoch": 27.38,
2534
+ "learning_rate": 0.00014390961025602855,
2535
+ "loss": 1.8113,
2536
+ "step": 4080
2537
+ },
2538
+ {
2539
+ "epoch": 27.45,
2540
+ "learning_rate": 0.00014254783804936686,
2541
+ "loss": 1.7796,
2542
+ "step": 4090
2543
+ },
2544
+ {
2545
+ "epoch": 27.52,
2546
+ "learning_rate": 0.00014118996530714664,
2547
+ "loss": 1.7518,
2548
+ "step": 4100
2549
+ },
2550
+ {
2551
+ "epoch": 27.58,
2552
+ "learning_rate": 0.00013983604130689192,
2553
+ "loss": 1.7064,
2554
+ "step": 4110
2555
+ },
2556
+ {
2557
+ "epoch": 27.65,
2558
+ "learning_rate": 0.0001384861151828255,
2559
+ "loss": 1.754,
2560
+ "step": 4120
2561
+ },
2562
+ {
2563
+ "epoch": 27.72,
2564
+ "learning_rate": 0.00013714023592408678,
2565
+ "loss": 1.7321,
2566
+ "step": 4130
2567
+ },
2568
+ {
2569
+ "epoch": 27.79,
2570
+ "learning_rate": 0.0001357984523729533,
2571
+ "loss": 1.787,
2572
+ "step": 4140
2573
+ },
2574
+ {
2575
+ "epoch": 27.85,
2576
+ "learning_rate": 0.00013446081322306813,
2577
+ "loss": 1.7389,
2578
+ "step": 4150
2579
+ },
2580
+ {
2581
+ "epoch": 27.92,
2582
+ "learning_rate": 0.00013312736701767347,
2583
+ "loss": 1.7504,
2584
+ "step": 4160
2585
+ },
2586
+ {
2587
+ "epoch": 27.99,
2588
+ "learning_rate": 0.00013179816214784826,
2589
+ "loss": 1.7239,
2590
+ "step": 4170
2591
+ },
2592
+ {
2593
+ "epoch": 28.05,
2594
+ "learning_rate": 0.00013047324685075263,
2595
+ "loss": 1.7251,
2596
+ "step": 4180
2597
+ },
2598
+ {
2599
+ "epoch": 28.12,
2600
+ "learning_rate": 0.00012915266920787672,
2601
+ "loss": 1.7394,
2602
+ "step": 4190
2603
+ },
2604
+ {
2605
+ "epoch": 28.19,
2606
+ "learning_rate": 0.00012783647714329649,
2607
+ "loss": 1.7623,
2608
+ "step": 4200
2609
+ },
2610
+ {
2611
+ "epoch": 28.26,
2612
+ "learning_rate": 0.00012652471842193415,
2613
+ "loss": 1.68,
2614
+ "step": 4210
2615
+ },
2616
+ {
2617
+ "epoch": 28.32,
2618
+ "learning_rate": 0.00012521744064782476,
2619
+ "loss": 1.7476,
2620
+ "step": 4220
2621
+ },
2622
+ {
2623
+ "epoch": 28.39,
2624
+ "learning_rate": 0.00012391469126238884,
2625
+ "loss": 1.7888,
2626
+ "step": 4230
2627
+ },
2628
+ {
2629
+ "epoch": 28.46,
2630
+ "learning_rate": 0.0001226165175427105,
2631
+ "loss": 1.8063,
2632
+ "step": 4240
2633
+ },
2634
+ {
2635
+ "epoch": 28.52,
2636
+ "learning_rate": 0.00012132296659982206,
2637
+ "loss": 1.7794,
2638
+ "step": 4250
2639
+ },
2640
+ {
2641
+ "epoch": 28.59,
2642
+ "learning_rate": 0.00012003408537699384,
2643
+ "loss": 1.7629,
2644
+ "step": 4260
2645
+ },
2646
+ {
2647
+ "epoch": 28.66,
2648
+ "learning_rate": 0.00011874992064803114,
2649
+ "loss": 1.7939,
2650
+ "step": 4270
2651
+ },
2652
+ {
2653
+ "epoch": 28.72,
2654
+ "learning_rate": 0.0001174705190155766,
2655
+ "loss": 1.7379,
2656
+ "step": 4280
2657
+ },
2658
+ {
2659
+ "epoch": 28.79,
2660
+ "learning_rate": 0.00011619592690941886,
2661
+ "loss": 1.7218,
2662
+ "step": 4290
2663
+ },
2664
+ {
2665
+ "epoch": 28.86,
2666
+ "learning_rate": 0.00011492619058480783,
2667
+ "loss": 1.7113,
2668
+ "step": 4300
2669
+ },
2670
+ {
2671
+ "epoch": 28.93,
2672
+ "learning_rate": 0.00011366135612077571,
2673
+ "loss": 1.7242,
2674
+ "step": 4310
2675
+ },
2676
+ {
2677
+ "epoch": 28.99,
2678
+ "learning_rate": 0.00011240146941846526,
2679
+ "loss": 1.7338,
2680
+ "step": 4320
2681
+ },
2682
+ {
2683
+ "epoch": 29.06,
2684
+ "learning_rate": 0.00011114657619946372,
2685
+ "loss": 1.67,
2686
+ "step": 4330
2687
+ },
2688
+ {
2689
+ "epoch": 29.13,
2690
+ "learning_rate": 0.00010989672200414375,
2691
+ "loss": 1.7605,
2692
+ "step": 4340
2693
+ },
2694
+ {
2695
+ "epoch": 29.19,
2696
+ "learning_rate": 0.00010865195219001028,
2697
+ "loss": 1.682,
2698
+ "step": 4350
2699
+ },
2700
+ {
2701
+ "epoch": 29.26,
2702
+ "learning_rate": 0.00010741231193005521,
2703
+ "loss": 1.6429,
2704
+ "step": 4360
2705
+ },
2706
+ {
2707
+ "epoch": 29.33,
2708
+ "learning_rate": 0.00010617784621111767,
2709
+ "loss": 1.6666,
2710
+ "step": 4370
2711
+ },
2712
+ {
2713
+ "epoch": 29.4,
2714
+ "learning_rate": 0.0001049485998322512,
2715
+ "loss": 1.6439,
2716
+ "step": 4380
2717
+ },
2718
+ {
2719
+ "epoch": 29.46,
2720
+ "learning_rate": 0.00010372461740309849,
2721
+ "loss": 1.6682,
2722
+ "step": 4390
2723
+ },
2724
+ {
2725
+ "epoch": 29.53,
2726
+ "learning_rate": 0.00010250594334227223,
2727
+ "loss": 1.8048,
2728
+ "step": 4400
2729
+ },
2730
+ {
2731
+ "epoch": 29.6,
2732
+ "learning_rate": 0.00010129262187574318,
2733
+ "loss": 1.7627,
2734
+ "step": 4410
2735
+ },
2736
+ {
2737
+ "epoch": 29.66,
2738
+ "learning_rate": 0.00010008469703523492,
2739
+ "loss": 1.8278,
2740
+ "step": 4420
2741
+ },
2742
+ {
2743
+ "epoch": 29.73,
2744
+ "learning_rate": 9.888221265662655e-05,
2745
+ "loss": 1.7661,
2746
+ "step": 4430
2747
+ },
2748
+ {
2749
+ "epoch": 29.8,
2750
+ "learning_rate": 9.768521237836131e-05,
2751
+ "loss": 1.7062,
2752
+ "step": 4440
2753
+ },
2754
+ {
2755
+ "epoch": 29.87,
2756
+ "learning_rate": 9.64937396398633e-05,
2757
+ "loss": 1.6505,
2758
+ "step": 4450
2759
+ },
2760
+ {
2761
+ "epoch": 29.93,
2762
+ "learning_rate": 9.530783767996057e-05,
2763
+ "loss": 1.688,
2764
+ "step": 4460
2765
+ },
2766
+ {
2767
+ "epoch": 30.0,
2768
+ "learning_rate": 9.412754953531663e-05,
2769
+ "loss": 1.6097,
2770
+ "step": 4470
2771
+ },
2772
+ {
2773
+ "epoch": 30.07,
2774
+ "learning_rate": 9.295291803886818e-05,
2775
+ "loss": 1.6393,
2776
+ "step": 4480
2777
+ },
2778
+ {
2779
+ "epoch": 30.13,
2780
+ "learning_rate": 9.178398581827086e-05,
2781
+ "loss": 1.6005,
2782
+ "step": 4490
2783
+ },
2784
+ {
2785
+ "epoch": 30.2,
2786
+ "learning_rate": 9.062079529435204e-05,
2787
+ "loss": 1.7324,
2788
+ "step": 4500
2789
+ },
2790
+ {
2791
+ "epoch": 30.2,
2792
+ "eval_accuracy": 0.7243114113807678,
2793
+ "eval_accuracy_top5": 0.8164724707603455,
2794
+ "eval_loss": 1.2510910034179688,
2795
+ "eval_runtime": 99.3635,
2796
+ "eval_samples_per_second": 38.072,
2797
+ "eval_steps_per_second": 4.76,
2798
+ "step": 4500
2799
+ },
2800
+ {
2801
+ "epoch": 30.27,
2802
+ "learning_rate": 8.946338867957182e-05,
2803
+ "loss": 1.6926,
2804
+ "step": 4510
2805
+ },
2806
+ {
2807
+ "epoch": 30.34,
2808
+ "learning_rate": 8.831180797649071e-05,
2809
+ "loss": 1.6246,
2810
+ "step": 4520
2811
+ },
2812
+ {
2813
+ "epoch": 30.4,
2814
+ "learning_rate": 8.716609497624564e-05,
2815
+ "loss": 1.7316,
2816
+ "step": 4530
2817
+ },
2818
+ {
2819
+ "epoch": 30.47,
2820
+ "learning_rate": 8.602629125703296e-05,
2821
+ "loss": 1.6764,
2822
+ "step": 4540
2823
+ },
2824
+ {
2825
+ "epoch": 30.54,
2826
+ "learning_rate": 8.489243818260004e-05,
2827
+ "loss": 1.664,
2828
+ "step": 4550
2829
+ },
2830
+ {
2831
+ "epoch": 30.6,
2832
+ "learning_rate": 8.376457690074384e-05,
2833
+ "loss": 1.6697,
2834
+ "step": 4560
2835
+ },
2836
+ {
2837
+ "epoch": 30.67,
2838
+ "learning_rate": 8.264274834181793e-05,
2839
+ "loss": 1.6584,
2840
+ "step": 4570
2841
+ },
2842
+ {
2843
+ "epoch": 30.74,
2844
+ "learning_rate": 8.152699321724652e-05,
2845
+ "loss": 1.6844,
2846
+ "step": 4580
2847
+ },
2848
+ {
2849
+ "epoch": 30.81,
2850
+ "learning_rate": 8.041735201804783e-05,
2851
+ "loss": 1.7346,
2852
+ "step": 4590
2853
+ },
2854
+ {
2855
+ "epoch": 30.87,
2856
+ "learning_rate": 7.931386501336418e-05,
2857
+ "loss": 1.7664,
2858
+ "step": 4600
2859
+ },
2860
+ {
2861
+ "epoch": 30.94,
2862
+ "learning_rate": 7.821657224900064e-05,
2863
+ "loss": 1.7337,
2864
+ "step": 4610
2865
+ },
2866
+ {
2867
+ "epoch": 31.01,
2868
+ "learning_rate": 7.712551354597187e-05,
2869
+ "loss": 1.5892,
2870
+ "step": 4620
2871
+ },
2872
+ {
2873
+ "epoch": 31.07,
2874
+ "learning_rate": 7.604072849905708e-05,
2875
+ "loss": 1.6845,
2876
+ "step": 4630
2877
+ },
2878
+ {
2879
+ "epoch": 31.14,
2880
+ "learning_rate": 7.49622564753627e-05,
2881
+ "loss": 1.6969,
2882
+ "step": 4640
2883
+ },
2884
+ {
2885
+ "epoch": 31.21,
2886
+ "learning_rate": 7.38901366128944e-05,
2887
+ "loss": 1.6825,
2888
+ "step": 4650
2889
+ },
2890
+ {
2891
+ "epoch": 31.28,
2892
+ "learning_rate": 7.282440781913619e-05,
2893
+ "loss": 1.6149,
2894
+ "step": 4660
2895
+ },
2896
+ {
2897
+ "epoch": 31.34,
2898
+ "learning_rate": 7.176510876963876e-05,
2899
+ "loss": 1.6943,
2900
+ "step": 4670
2901
+ },
2902
+ {
2903
+ "epoch": 31.41,
2904
+ "learning_rate": 7.071227790661597e-05,
2905
+ "loss": 1.7003,
2906
+ "step": 4680
2907
+ },
2908
+ {
2909
+ "epoch": 31.48,
2910
+ "learning_rate": 6.966595343754936e-05,
2911
+ "loss": 1.754,
2912
+ "step": 4690
2913
+ },
2914
+ {
2915
+ "epoch": 31.54,
2916
+ "learning_rate": 6.862617333380214e-05,
2917
+ "loss": 1.7057,
2918
+ "step": 4700
2919
+ },
2920
+ {
2921
+ "epoch": 31.61,
2922
+ "learning_rate": 6.759297532924086e-05,
2923
+ "loss": 1.6623,
2924
+ "step": 4710
2925
+ },
2926
+ {
2927
+ "epoch": 31.68,
2928
+ "learning_rate": 6.656639691886629e-05,
2929
+ "loss": 1.5663,
2930
+ "step": 4720
2931
+ },
2932
+ {
2933
+ "epoch": 31.74,
2934
+ "learning_rate": 6.55464753574522e-05,
2935
+ "loss": 1.62,
2936
+ "step": 4730
2937
+ },
2938
+ {
2939
+ "epoch": 31.81,
2940
+ "learning_rate": 6.453324765819404e-05,
2941
+ "loss": 1.6554,
2942
+ "step": 4740
2943
+ },
2944
+ {
2945
+ "epoch": 31.88,
2946
+ "learning_rate": 6.352675059136531e-05,
2947
+ "loss": 1.5982,
2948
+ "step": 4750
2949
+ },
2950
+ {
2951
+ "epoch": 31.95,
2952
+ "learning_rate": 6.252702068298338e-05,
2953
+ "loss": 1.681,
2954
+ "step": 4760
2955
+ },
2956
+ {
2957
+ "epoch": 32.01,
2958
+ "learning_rate": 6.153409421348358e-05,
2959
+ "loss": 1.6682,
2960
+ "step": 4770
2961
+ },
2962
+ {
2963
+ "epoch": 32.08,
2964
+ "learning_rate": 6.054800721640305e-05,
2965
+ "loss": 1.6059,
2966
+ "step": 4780
2967
+ },
2968
+ {
2969
+ "epoch": 32.15,
2970
+ "learning_rate": 5.956879547707275e-05,
2971
+ "loss": 1.661,
2972
+ "step": 4790
2973
+ },
2974
+ {
2975
+ "epoch": 32.21,
2976
+ "learning_rate": 5.8596494531319045e-05,
2977
+ "loss": 1.6645,
2978
+ "step": 4800
2979
+ },
2980
+ {
2981
+ "epoch": 32.28,
2982
+ "learning_rate": 5.763113966417369e-05,
2983
+ "loss": 1.6481,
2984
+ "step": 4810
2985
+ },
2986
+ {
2987
+ "epoch": 32.35,
2988
+ "learning_rate": 5.667276590859385e-05,
2989
+ "loss": 1.6029,
2990
+ "step": 4820
2991
+ },
2992
+ {
2993
+ "epoch": 32.42,
2994
+ "learning_rate": 5.572140804419049e-05,
2995
+ "loss": 1.6251,
2996
+ "step": 4830
2997
+ },
2998
+ {
2999
+ "epoch": 32.48,
3000
+ "learning_rate": 5.477710059596599e-05,
3001
+ "loss": 1.676,
3002
+ "step": 4840
3003
+ },
3004
+ {
3005
+ "epoch": 32.55,
3006
+ "learning_rate": 5.3839877833061684e-05,
3007
+ "loss": 1.6486,
3008
+ "step": 4850
3009
+ },
3010
+ {
3011
+ "epoch": 32.62,
3012
+ "learning_rate": 5.2909773767513934e-05,
3013
+ "loss": 1.6394,
3014
+ "step": 4860
3015
+ },
3016
+ {
3017
+ "epoch": 32.68,
3018
+ "learning_rate": 5.198682215301989e-05,
3019
+ "loss": 1.5951,
3020
+ "step": 4870
3021
+ },
3022
+ {
3023
+ "epoch": 32.75,
3024
+ "learning_rate": 5.1071056483712435e-05,
3025
+ "loss": 1.6508,
3026
+ "step": 4880
3027
+ },
3028
+ {
3029
+ "epoch": 32.82,
3030
+ "learning_rate": 5.016250999294497e-05,
3031
+ "loss": 1.615,
3032
+ "step": 4890
3033
+ },
3034
+ {
3035
+ "epoch": 32.89,
3036
+ "learning_rate": 4.9261215652085105e-05,
3037
+ "loss": 1.6343,
3038
+ "step": 4900
3039
+ },
3040
+ {
3041
+ "epoch": 32.95,
3042
+ "learning_rate": 4.8367206169318305e-05,
3043
+ "loss": 1.608,
3044
+ "step": 4910
3045
+ },
3046
+ {
3047
+ "epoch": 33.02,
3048
+ "learning_rate": 4.7480513988460625e-05,
3049
+ "loss": 1.5985,
3050
+ "step": 4920
3051
+ },
3052
+ {
3053
+ "epoch": 33.09,
3054
+ "learning_rate": 4.660117128778163e-05,
3055
+ "loss": 1.6053,
3056
+ "step": 4930
3057
+ },
3058
+ {
3059
+ "epoch": 33.15,
3060
+ "learning_rate": 4.572920997883648e-05,
3061
+ "loss": 1.6875,
3062
+ "step": 4940
3063
+ },
3064
+ {
3065
+ "epoch": 33.22,
3066
+ "learning_rate": 4.486466170530798e-05,
3067
+ "loss": 1.6692,
3068
+ "step": 4950
3069
+ },
3070
+ {
3071
+ "epoch": 33.29,
3072
+ "learning_rate": 4.4007557841857865e-05,
3073
+ "loss": 1.7254,
3074
+ "step": 4960
3075
+ },
3076
+ {
3077
+ "epoch": 33.36,
3078
+ "learning_rate": 4.315792949298869e-05,
3079
+ "loss": 1.6377,
3080
+ "step": 4970
3081
+ },
3082
+ {
3083
+ "epoch": 33.42,
3084
+ "learning_rate": 4.231580749191474e-05,
3085
+ "loss": 1.6221,
3086
+ "step": 4980
3087
+ },
3088
+ {
3089
+ "epoch": 33.49,
3090
+ "learning_rate": 4.148122239944316e-05,
3091
+ "loss": 1.7078,
3092
+ "step": 4990
3093
+ },
3094
+ {
3095
+ "epoch": 33.56,
3096
+ "learning_rate": 4.0654204502864886e-05,
3097
+ "loss": 1.5947,
3098
+ "step": 5000
3099
+ },
3100
+ {
3101
+ "epoch": 33.56,
3102
+ "eval_accuracy": 0.7288135886192322,
3103
+ "eval_accuracy_top5": 0.8172669410705566,
3104
+ "eval_loss": 1.2728419303894043,
3105
+ "eval_runtime": 100.0466,
3106
+ "eval_samples_per_second": 37.812,
3107
+ "eval_steps_per_second": 4.728,
3108
+ "step": 5000
3109
+ },
3110
+ {
3111
+ "epoch": 33.62,
3112
+ "learning_rate": 3.983478381485558e-05,
3113
+ "loss": 1.6253,
3114
+ "step": 5010
3115
+ },
3116
+ {
3117
+ "epoch": 33.69,
3118
+ "learning_rate": 3.902299007238627e-05,
3119
+ "loss": 1.6119,
3120
+ "step": 5020
3121
+ },
3122
+ {
3123
+ "epoch": 33.76,
3124
+ "learning_rate": 3.8218852735644404e-05,
3125
+ "loss": 1.5703,
3126
+ "step": 5030
3127
+ },
3128
+ {
3129
+ "epoch": 33.83,
3130
+ "learning_rate": 3.7422400986964724e-05,
3131
+ "loss": 1.6635,
3132
+ "step": 5040
3133
+ },
3134
+ {
3135
+ "epoch": 33.89,
3136
+ "learning_rate": 3.6633663729770004e-05,
3137
+ "loss": 1.6116,
3138
+ "step": 5050
3139
+ },
3140
+ {
3141
+ "epoch": 33.96,
3142
+ "learning_rate": 3.585266958752248e-05,
3143
+ "loss": 1.5966,
3144
+ "step": 5060
3145
+ },
3146
+ {
3147
+ "epoch": 34.03,
3148
+ "learning_rate": 3.507944690268469e-05,
3149
+ "loss": 1.656,
3150
+ "step": 5070
3151
+ },
3152
+ {
3153
+ "epoch": 34.09,
3154
+ "learning_rate": 3.4314023735691286e-05,
3155
+ "loss": 1.6135,
3156
+ "step": 5080
3157
+ },
3158
+ {
3159
+ "epoch": 34.16,
3160
+ "learning_rate": 3.355642786393051e-05,
3161
+ "loss": 1.6592,
3162
+ "step": 5090
3163
+ },
3164
+ {
3165
+ "epoch": 34.23,
3166
+ "learning_rate": 3.2806686780736336e-05,
3167
+ "loss": 1.5753,
3168
+ "step": 5100
3169
+ },
3170
+ {
3171
+ "epoch": 34.3,
3172
+ "learning_rate": 3.2064827694390345e-05,
3173
+ "loss": 1.6342,
3174
+ "step": 5110
3175
+ },
3176
+ {
3177
+ "epoch": 34.36,
3178
+ "learning_rate": 3.1330877527134785e-05,
3179
+ "loss": 1.6055,
3180
+ "step": 5120
3181
+ },
3182
+ {
3183
+ "epoch": 34.43,
3184
+ "learning_rate": 3.060486291419531e-05,
3185
+ "loss": 1.6282,
3186
+ "step": 5130
3187
+ },
3188
+ {
3189
+ "epoch": 34.5,
3190
+ "learning_rate": 2.9886810202814447e-05,
3191
+ "loss": 1.6759,
3192
+ "step": 5140
3193
+ },
3194
+ {
3195
+ "epoch": 34.56,
3196
+ "learning_rate": 2.917674545129531e-05,
3197
+ "loss": 1.6517,
3198
+ "step": 5150
3199
+ },
3200
+ {
3201
+ "epoch": 34.63,
3202
+ "learning_rate": 2.847469442805614e-05,
3203
+ "loss": 1.5875,
3204
+ "step": 5160
3205
+ },
3206
+ {
3207
+ "epoch": 34.7,
3208
+ "learning_rate": 2.7780682610695136e-05,
3209
+ "loss": 1.6003,
3210
+ "step": 5170
3211
+ },
3212
+ {
3213
+ "epoch": 34.77,
3214
+ "learning_rate": 2.7094735185065778e-05,
3215
+ "loss": 1.5555,
3216
+ "step": 5180
3217
+ },
3218
+ {
3219
+ "epoch": 34.83,
3220
+ "learning_rate": 2.6416877044362685e-05,
3221
+ "loss": 1.6044,
3222
+ "step": 5190
3223
+ },
3224
+ {
3225
+ "epoch": 34.9,
3226
+ "learning_rate": 2.5747132788218663e-05,
3227
+ "loss": 1.6296,
3228
+ "step": 5200
3229
+ },
3230
+ {
3231
+ "epoch": 34.97,
3232
+ "learning_rate": 2.508552672181158e-05,
3233
+ "loss": 1.6314,
3234
+ "step": 5210
3235
+ },
3236
+ {
3237
+ "epoch": 35.03,
3238
+ "learning_rate": 2.4432082854982524e-05,
3239
+ "loss": 1.6526,
3240
+ "step": 5220
3241
+ },
3242
+ {
3243
+ "epoch": 35.1,
3244
+ "learning_rate": 2.3786824901364357e-05,
3245
+ "loss": 1.6304,
3246
+ "step": 5230
3247
+ },
3248
+ {
3249
+ "epoch": 35.17,
3250
+ "learning_rate": 2.3149776277521266e-05,
3251
+ "loss": 1.5172,
3252
+ "step": 5240
3253
+ },
3254
+ {
3255
+ "epoch": 35.23,
3256
+ "learning_rate": 2.2520960102098892e-05,
3257
+ "loss": 1.606,
3258
+ "step": 5250
3259
+ },
3260
+ {
3261
+ "epoch": 35.3,
3262
+ "learning_rate": 2.190039919498543e-05,
3263
+ "loss": 1.5952,
3264
+ "step": 5260
3265
+ },
3266
+ {
3267
+ "epoch": 35.37,
3268
+ "learning_rate": 2.1288116076483288e-05,
3269
+ "loss": 1.5936,
3270
+ "step": 5270
3271
+ },
3272
+ {
3273
+ "epoch": 35.44,
3274
+ "learning_rate": 2.0684132966492103e-05,
3275
+ "loss": 1.6147,
3276
+ "step": 5280
3277
+ },
3278
+ {
3279
+ "epoch": 35.5,
3280
+ "learning_rate": 2.008847178370221e-05,
3281
+ "loss": 1.6177,
3282
+ "step": 5290
3283
+ },
3284
+ {
3285
+ "epoch": 35.57,
3286
+ "learning_rate": 1.9501154144799137e-05,
3287
+ "loss": 1.6251,
3288
+ "step": 5300
3289
+ },
3290
+ {
3291
+ "epoch": 35.64,
3292
+ "learning_rate": 1.8922201363679338e-05,
3293
+ "loss": 1.6342,
3294
+ "step": 5310
3295
+ },
3296
+ {
3297
+ "epoch": 35.7,
3298
+ "learning_rate": 1.8351634450676502e-05,
3299
+ "loss": 1.6048,
3300
+ "step": 5320
3301
+ },
3302
+ {
3303
+ "epoch": 35.77,
3304
+ "learning_rate": 1.7789474111799318e-05,
3305
+ "loss": 1.5749,
3306
+ "step": 5330
3307
+ },
3308
+ {
3309
+ "epoch": 35.84,
3310
+ "learning_rate": 1.7235740747979733e-05,
3311
+ "loss": 1.6196,
3312
+ "step": 5340
3313
+ },
3314
+ {
3315
+ "epoch": 35.91,
3316
+ "learning_rate": 1.6690454454332843e-05,
3317
+ "loss": 1.5682,
3318
+ "step": 5350
3319
+ },
3320
+ {
3321
+ "epoch": 35.97,
3322
+ "learning_rate": 1.6153635019427598e-05,
3323
+ "loss": 1.6569,
3324
+ "step": 5360
3325
+ },
3326
+ {
3327
+ "epoch": 36.04,
3328
+ "learning_rate": 1.5625301924568626e-05,
3329
+ "loss": 1.6054,
3330
+ "step": 5370
3331
+ },
3332
+ {
3333
+ "epoch": 36.11,
3334
+ "learning_rate": 1.5105474343089205e-05,
3335
+ "loss": 1.6504,
3336
+ "step": 5380
3337
+ },
3338
+ {
3339
+ "epoch": 36.17,
3340
+ "learning_rate": 1.459417113965561e-05,
3341
+ "loss": 1.606,
3342
+ "step": 5390
3343
+ },
3344
+ {
3345
+ "epoch": 36.24,
3346
+ "learning_rate": 1.4091410869582266e-05,
3347
+ "loss": 1.627,
3348
+ "step": 5400
3349
+ },
3350
+ {
3351
+ "epoch": 36.31,
3352
+ "learning_rate": 1.3597211778158675e-05,
3353
+ "loss": 1.5914,
3354
+ "step": 5410
3355
+ },
3356
+ {
3357
+ "epoch": 36.38,
3358
+ "learning_rate": 1.3111591799987083e-05,
3359
+ "loss": 1.6279,
3360
+ "step": 5420
3361
+ },
3362
+ {
3363
+ "epoch": 36.44,
3364
+ "learning_rate": 1.2634568558331644e-05,
3365
+ "loss": 1.6583,
3366
+ "step": 5430
3367
+ },
3368
+ {
3369
+ "epoch": 36.51,
3370
+ "learning_rate": 1.216615936447893e-05,
3371
+ "loss": 1.567,
3372
+ "step": 5440
3373
+ },
3374
+ {
3375
+ "epoch": 36.58,
3376
+ "learning_rate": 1.1706381217109735e-05,
3377
+ "loss": 1.6408,
3378
+ "step": 5450
3379
+ },
3380
+ {
3381
+ "epoch": 36.64,
3382
+ "learning_rate": 1.1255250801681994e-05,
3383
+ "loss": 1.5914,
3384
+ "step": 5460
3385
+ },
3386
+ {
3387
+ "epoch": 36.71,
3388
+ "learning_rate": 1.0812784489825506e-05,
3389
+ "loss": 1.5906,
3390
+ "step": 5470
3391
+ },
3392
+ {
3393
+ "epoch": 36.78,
3394
+ "learning_rate": 1.0378998338747669e-05,
3395
+ "loss": 1.541,
3396
+ "step": 5480
3397
+ },
3398
+ {
3399
+ "epoch": 36.85,
3400
+ "learning_rate": 9.953908090650804e-06,
3401
+ "loss": 1.6335,
3402
+ "step": 5490
3403
+ },
3404
+ {
3405
+ "epoch": 36.91,
3406
+ "learning_rate": 9.53752917216083e-06,
3407
+ "loss": 1.6565,
3408
+ "step": 5500
3409
+ },
3410
+ {
3411
+ "epoch": 36.91,
3412
+ "eval_accuracy": 0.7513241767883301,
3413
+ "eval_accuracy_top5": 0.8336864113807678,
3414
+ "eval_loss": 1.1887362003326416,
3415
+ "eval_runtime": 101.4114,
3416
+ "eval_samples_per_second": 37.304,
3417
+ "eval_steps_per_second": 4.664,
3418
+ "step": 5500
3419
+ }
3420
+ ],
3421
+ "max_steps": 5960,
3422
+ "num_train_epochs": 40,
3423
+ "total_flos": 4.3900198664955494e+20,
3424
+ "trial_name": null,
3425
+ "trial_params": null
3426
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a4e980947c7fdde72550b59634bd952c86ddfeb9e69c639aee0d5c44812cf0a
3
+ size 3451