Upload 10 files
Browse files- config.json +52 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- spiece.model +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- trainer_state.json +2193 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "albert-base-v2",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"AlbertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0,
|
| 7 |
+
"bos_token_id": 2,
|
| 8 |
+
"classifier_dropout_prob": 0.1,
|
| 9 |
+
"down_scale_factor": 1,
|
| 10 |
+
"embedding_size": 128,
|
| 11 |
+
"eos_token_id": 3,
|
| 12 |
+
"gap_size": 0,
|
| 13 |
+
"hidden_act": "gelu_new",
|
| 14 |
+
"hidden_dropout_prob": 0,
|
| 15 |
+
"hidden_size": 768,
|
| 16 |
+
"id2label": {
|
| 17 |
+
"0": "LABEL_0",
|
| 18 |
+
"1": "LABEL_1",
|
| 19 |
+
"2": "LABEL_2",
|
| 20 |
+
"3": "LABEL_3",
|
| 21 |
+
"4": "LABEL_4",
|
| 22 |
+
"5": "LABEL_5",
|
| 23 |
+
"6": "LABEL_6"
|
| 24 |
+
},
|
| 25 |
+
"initializer_range": 0.02,
|
| 26 |
+
"inner_group_num": 1,
|
| 27 |
+
"intermediate_size": 3072,
|
| 28 |
+
"label2id": {
|
| 29 |
+
"LABEL_0": 0,
|
| 30 |
+
"LABEL_1": 1,
|
| 31 |
+
"LABEL_2": 2,
|
| 32 |
+
"LABEL_3": 3,
|
| 33 |
+
"LABEL_4": 4,
|
| 34 |
+
"LABEL_5": 5,
|
| 35 |
+
"LABEL_6": 6
|
| 36 |
+
},
|
| 37 |
+
"layer_norm_eps": 1e-12,
|
| 38 |
+
"max_position_embeddings": 512,
|
| 39 |
+
"model_type": "albert",
|
| 40 |
+
"net_structure_type": 0,
|
| 41 |
+
"num_attention_heads": 12,
|
| 42 |
+
"num_hidden_groups": 1,
|
| 43 |
+
"num_hidden_layers": 12,
|
| 44 |
+
"num_memory_blocks": 0,
|
| 45 |
+
"pad_token_id": 0,
|
| 46 |
+
"position_embedding_type": "absolute",
|
| 47 |
+
"problem_type": "single_label_classification",
|
| 48 |
+
"torch_dtype": "float32",
|
| 49 |
+
"transformers_version": "4.40.2",
|
| 50 |
+
"type_vocab_size": 2,
|
| 51 |
+
"vocab_size": 30000
|
| 52 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e482016854d986d9502fab435fd351b7524150fdb48baa84d2046d37b28047bd
|
| 3 |
+
size 46759292
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:704f791ee5f506d1446755aba52655f550919e98e43e56ccdc85c0193627fb08
|
| 3 |
+
size 93534733
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eaf7361eaba266493774b86d64b97ccfa7c8e3649ff56cfcc431422b4296fa35
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c203a76759ba1afd42aaa905214564ee6ba1e700114d1c8f8020faabdd31632
|
| 3 |
+
size 1064
|
spiece.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fefb02b667a6c5c2fe27602d28e5fb3428f66ab89c7d6f388e7c8d44a02d0336
|
| 3 |
+
size 760289
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"model_max_length": 512}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2193 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9315417408943176,
|
| 3 |
+
"best_model_checkpoint": "./albert_sentiment_model\\checkpoint-2000",
|
| 4 |
+
"epoch": 2.3094688221709005,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.007698229407236336,
|
| 13 |
+
"grad_norm": 25.626720428466797,
|
| 14 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 15 |
+
"loss": 1.9667,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.015396458814472672,
|
| 20 |
+
"grad_norm": 35.37693405151367,
|
| 21 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 22 |
+
"loss": 2.0464,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.023094688221709007,
|
| 27 |
+
"grad_norm": 22.576478958129883,
|
| 28 |
+
"learning_rate": 3e-06,
|
| 29 |
+
"loss": 1.9565,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.030792917628945343,
|
| 34 |
+
"grad_norm": 21.228620529174805,
|
| 35 |
+
"learning_rate": 4.000000000000001e-06,
|
| 36 |
+
"loss": 1.9974,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.03849114703618168,
|
| 41 |
+
"grad_norm": 27.480052947998047,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 1.9431,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.046189376443418015,
|
| 48 |
+
"grad_norm": 27.75535774230957,
|
| 49 |
+
"learning_rate": 6e-06,
|
| 50 |
+
"loss": 1.8815,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05388760585065435,
|
| 55 |
+
"grad_norm": 17.684391021728516,
|
| 56 |
+
"learning_rate": 7.000000000000001e-06,
|
| 57 |
+
"loss": 1.8753,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.061585835257890686,
|
| 62 |
+
"grad_norm": 21.62338638305664,
|
| 63 |
+
"learning_rate": 8.000000000000001e-06,
|
| 64 |
+
"loss": 1.8629,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.06928406466512702,
|
| 69 |
+
"grad_norm": 43.68171691894531,
|
| 70 |
+
"learning_rate": 9e-06,
|
| 71 |
+
"loss": 1.8003,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.07698229407236336,
|
| 76 |
+
"grad_norm": 22.630041122436523,
|
| 77 |
+
"learning_rate": 1e-05,
|
| 78 |
+
"loss": 1.7636,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.08468052347959969,
|
| 83 |
+
"grad_norm": 30.946426391601562,
|
| 84 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 85 |
+
"loss": 1.7074,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.09237875288683603,
|
| 90 |
+
"grad_norm": 30.420377731323242,
|
| 91 |
+
"learning_rate": 1.2e-05,
|
| 92 |
+
"loss": 1.6716,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.10007698229407236,
|
| 97 |
+
"grad_norm": 48.35504150390625,
|
| 98 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 99 |
+
"loss": 1.4465,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.1077752117013087,
|
| 104 |
+
"grad_norm": 51.120361328125,
|
| 105 |
+
"learning_rate": 1.4000000000000001e-05,
|
| 106 |
+
"loss": 1.4337,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.11547344110854503,
|
| 111 |
+
"grad_norm": 29.794086456298828,
|
| 112 |
+
"learning_rate": 1.5e-05,
|
| 113 |
+
"loss": 1.3535,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.12317167051578137,
|
| 118 |
+
"grad_norm": 34.01353454589844,
|
| 119 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 120 |
+
"loss": 1.3233,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1308698999230177,
|
| 125 |
+
"grad_norm": 47.1700439453125,
|
| 126 |
+
"learning_rate": 1.7000000000000003e-05,
|
| 127 |
+
"loss": 1.342,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.13856812933025403,
|
| 132 |
+
"grad_norm": 34.01679611206055,
|
| 133 |
+
"learning_rate": 1.8e-05,
|
| 134 |
+
"loss": 1.4129,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1462663587374904,
|
| 139 |
+
"grad_norm": 27.47703742980957,
|
| 140 |
+
"learning_rate": 1.9e-05,
|
| 141 |
+
"loss": 1.3769,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.15396458814472672,
|
| 146 |
+
"grad_norm": 60.44585037231445,
|
| 147 |
+
"learning_rate": 2e-05,
|
| 148 |
+
"loss": 1.3133,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.16166281755196305,
|
| 153 |
+
"grad_norm": 23.065670013427734,
|
| 154 |
+
"learning_rate": 2.1e-05,
|
| 155 |
+
"loss": 1.2592,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.16936104695919937,
|
| 160 |
+
"grad_norm": 70.744873046875,
|
| 161 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 162 |
+
"loss": 1.3765,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.17705927636643573,
|
| 167 |
+
"grad_norm": 47.14573287963867,
|
| 168 |
+
"learning_rate": 2.3000000000000003e-05,
|
| 169 |
+
"loss": 1.3629,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.18475750577367206,
|
| 174 |
+
"grad_norm": 26.078250885009766,
|
| 175 |
+
"learning_rate": 2.4e-05,
|
| 176 |
+
"loss": 1.1865,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1924557351809084,
|
| 181 |
+
"grad_norm": 63.99885940551758,
|
| 182 |
+
"learning_rate": 2.5e-05,
|
| 183 |
+
"loss": 1.3399,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.20015396458814472,
|
| 188 |
+
"grad_norm": 24.421600341796875,
|
| 189 |
+
"learning_rate": 2.6000000000000002e-05,
|
| 190 |
+
"loss": 1.1854,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.20785219399538107,
|
| 195 |
+
"grad_norm": 16.544498443603516,
|
| 196 |
+
"learning_rate": 2.7000000000000002e-05,
|
| 197 |
+
"loss": 1.2511,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.2155504234026174,
|
| 202 |
+
"grad_norm": 52.77935791015625,
|
| 203 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 204 |
+
"loss": 1.2485,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.22324865280985373,
|
| 209 |
+
"grad_norm": 68.4757080078125,
|
| 210 |
+
"learning_rate": 2.9e-05,
|
| 211 |
+
"loss": 1.0838,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.23094688221709006,
|
| 216 |
+
"grad_norm": 26.108631134033203,
|
| 217 |
+
"learning_rate": 3e-05,
|
| 218 |
+
"loss": 1.3167,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.23864511162432642,
|
| 223 |
+
"grad_norm": 16.388147354125977,
|
| 224 |
+
"learning_rate": 3.1e-05,
|
| 225 |
+
"loss": 1.0942,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.24634334103156275,
|
| 230 |
+
"grad_norm": 27.521499633789062,
|
| 231 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 232 |
+
"loss": 1.3823,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2540415704387991,
|
| 237 |
+
"grad_norm": 18.420778274536133,
|
| 238 |
+
"learning_rate": 3.3e-05,
|
| 239 |
+
"loss": 1.1258,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.2617397998460354,
|
| 244 |
+
"grad_norm": 16.579042434692383,
|
| 245 |
+
"learning_rate": 3.4000000000000007e-05,
|
| 246 |
+
"loss": 1.159,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.26943802925327176,
|
| 251 |
+
"grad_norm": 48.20399856567383,
|
| 252 |
+
"learning_rate": 3.5e-05,
|
| 253 |
+
"loss": 1.1556,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.27713625866050806,
|
| 258 |
+
"grad_norm": 12.147098541259766,
|
| 259 |
+
"learning_rate": 3.6e-05,
|
| 260 |
+
"loss": 1.1309,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2848344880677444,
|
| 265 |
+
"grad_norm": 52.27555847167969,
|
| 266 |
+
"learning_rate": 3.7e-05,
|
| 267 |
+
"loss": 1.3251,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.2925327174749808,
|
| 272 |
+
"grad_norm": 42.3845100402832,
|
| 273 |
+
"learning_rate": 3.8e-05,
|
| 274 |
+
"loss": 1.3005,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3002309468822171,
|
| 279 |
+
"grad_norm": 29.475187301635742,
|
| 280 |
+
"learning_rate": 3.9000000000000006e-05,
|
| 281 |
+
"loss": 1.2515,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.30792917628945343,
|
| 286 |
+
"grad_norm": 60.33949279785156,
|
| 287 |
+
"learning_rate": 4e-05,
|
| 288 |
+
"loss": 1.2975,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.3156274056966898,
|
| 293 |
+
"grad_norm": 47.650211334228516,
|
| 294 |
+
"learning_rate": 4.1e-05,
|
| 295 |
+
"loss": 1.2688,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.3233256351039261,
|
| 300 |
+
"grad_norm": 20.859683990478516,
|
| 301 |
+
"learning_rate": 4.2e-05,
|
| 302 |
+
"loss": 1.2035,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.33102386451116245,
|
| 307 |
+
"grad_norm": 16.41029930114746,
|
| 308 |
+
"learning_rate": 4.3e-05,
|
| 309 |
+
"loss": 1.1203,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.33872209391839875,
|
| 314 |
+
"grad_norm": 12.632010459899902,
|
| 315 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 316 |
+
"loss": 1.1197,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.3464203233256351,
|
| 321 |
+
"grad_norm": 36.34510803222656,
|
| 322 |
+
"learning_rate": 4.5e-05,
|
| 323 |
+
"loss": 1.2378,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.35411855273287146,
|
| 328 |
+
"grad_norm": 52.647926330566406,
|
| 329 |
+
"learning_rate": 4.600000000000001e-05,
|
| 330 |
+
"loss": 1.2022,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.36181678214010776,
|
| 335 |
+
"grad_norm": 88.51087188720703,
|
| 336 |
+
"learning_rate": 4.7e-05,
|
| 337 |
+
"loss": 1.0755,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3695150115473441,
|
| 342 |
+
"grad_norm": 27.021543502807617,
|
| 343 |
+
"learning_rate": 4.8e-05,
|
| 344 |
+
"loss": 1.2264,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.3772132409545804,
|
| 349 |
+
"grad_norm": 13.925777435302734,
|
| 350 |
+
"learning_rate": 4.9e-05,
|
| 351 |
+
"loss": 1.0287,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.3849114703618168,
|
| 356 |
+
"grad_norm": 17.583614349365234,
|
| 357 |
+
"learning_rate": 5e-05,
|
| 358 |
+
"loss": 0.979,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.3849114703618168,
|
| 363 |
+
"eval_accuracy": 0.5254041570438799,
|
| 364 |
+
"eval_f1": 0.4643762968149376,
|
| 365 |
+
"eval_loss": 1.1208206415176392,
|
| 366 |
+
"eval_precision": 0.47090986837941123,
|
| 367 |
+
"eval_recall": 0.5254041570438799,
|
| 368 |
+
"eval_runtime": 7.9109,
|
| 369 |
+
"eval_samples_per_second": 328.406,
|
| 370 |
+
"eval_steps_per_second": 41.082,
|
| 371 |
+
"step": 500
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.39260969976905313,
|
| 375 |
+
"grad_norm": 29.161548614501953,
|
| 376 |
+
"learning_rate": 4.998700285937094e-05,
|
| 377 |
+
"loss": 1.1457,
|
| 378 |
+
"step": 510
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 0.40030792917628943,
|
| 382 |
+
"grad_norm": 45.45001220703125,
|
| 383 |
+
"learning_rate": 4.997400571874188e-05,
|
| 384 |
+
"loss": 1.4028,
|
| 385 |
+
"step": 520
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 0.4080061585835258,
|
| 389 |
+
"grad_norm": 20.393016815185547,
|
| 390 |
+
"learning_rate": 4.9961008578112814e-05,
|
| 391 |
+
"loss": 1.3193,
|
| 392 |
+
"step": 530
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.41570438799076215,
|
| 396 |
+
"grad_norm": 9.973402976989746,
|
| 397 |
+
"learning_rate": 4.9948011437483756e-05,
|
| 398 |
+
"loss": 1.088,
|
| 399 |
+
"step": 540
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.42340261739799845,
|
| 403 |
+
"grad_norm": 41.007110595703125,
|
| 404 |
+
"learning_rate": 4.993501429685469e-05,
|
| 405 |
+
"loss": 1.3676,
|
| 406 |
+
"step": 550
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 0.4311008468052348,
|
| 410 |
+
"grad_norm": 13.304949760437012,
|
| 411 |
+
"learning_rate": 4.992201715622563e-05,
|
| 412 |
+
"loss": 1.3313,
|
| 413 |
+
"step": 560
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.4387990762124711,
|
| 417 |
+
"grad_norm": 10.015753746032715,
|
| 418 |
+
"learning_rate": 4.990902001559657e-05,
|
| 419 |
+
"loss": 1.0298,
|
| 420 |
+
"step": 570
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 0.44649730561970746,
|
| 424 |
+
"grad_norm": 67.66326904296875,
|
| 425 |
+
"learning_rate": 4.989602287496751e-05,
|
| 426 |
+
"loss": 1.0524,
|
| 427 |
+
"step": 580
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 0.4541955350269438,
|
| 431 |
+
"grad_norm": 28.423460006713867,
|
| 432 |
+
"learning_rate": 4.9883025734338444e-05,
|
| 433 |
+
"loss": 1.1959,
|
| 434 |
+
"step": 590
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.4618937644341801,
|
| 438 |
+
"grad_norm": 18.55489158630371,
|
| 439 |
+
"learning_rate": 4.9870028593709386e-05,
|
| 440 |
+
"loss": 1.0586,
|
| 441 |
+
"step": 600
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.4695919938414165,
|
| 445 |
+
"grad_norm": 36.054500579833984,
|
| 446 |
+
"learning_rate": 4.985703145308033e-05,
|
| 447 |
+
"loss": 1.2009,
|
| 448 |
+
"step": 610
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 0.47729022324865283,
|
| 452 |
+
"grad_norm": 18.236879348754883,
|
| 453 |
+
"learning_rate": 4.984403431245126e-05,
|
| 454 |
+
"loss": 1.1012,
|
| 455 |
+
"step": 620
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 0.48498845265588914,
|
| 459 |
+
"grad_norm": 20.71109962463379,
|
| 460 |
+
"learning_rate": 4.9831037171822204e-05,
|
| 461 |
+
"loss": 1.1486,
|
| 462 |
+
"step": 630
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 0.4926866820631255,
|
| 466 |
+
"grad_norm": 28.097261428833008,
|
| 467 |
+
"learning_rate": 4.981804003119314e-05,
|
| 468 |
+
"loss": 1.1645,
|
| 469 |
+
"step": 640
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 0.5003849114703618,
|
| 473 |
+
"grad_norm": 20.31618309020996,
|
| 474 |
+
"learning_rate": 4.980504289056408e-05,
|
| 475 |
+
"loss": 1.1256,
|
| 476 |
+
"step": 650
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.5080831408775982,
|
| 480 |
+
"grad_norm": 126.53520202636719,
|
| 481 |
+
"learning_rate": 4.9792045749935016e-05,
|
| 482 |
+
"loss": 1.4823,
|
| 483 |
+
"step": 660
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.5157813702848345,
|
| 487 |
+
"grad_norm": 14.448310852050781,
|
| 488 |
+
"learning_rate": 4.977904860930596e-05,
|
| 489 |
+
"loss": 1.2872,
|
| 490 |
+
"step": 670
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.5234795996920708,
|
| 494 |
+
"grad_norm": 16.949066162109375,
|
| 495 |
+
"learning_rate": 4.976605146867689e-05,
|
| 496 |
+
"loss": 1.2012,
|
| 497 |
+
"step": 680
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 0.5311778290993071,
|
| 501 |
+
"grad_norm": 14.372076034545898,
|
| 502 |
+
"learning_rate": 4.9753054328047835e-05,
|
| 503 |
+
"loss": 1.1928,
|
| 504 |
+
"step": 690
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.5388760585065435,
|
| 508 |
+
"grad_norm": 76.28020477294922,
|
| 509 |
+
"learning_rate": 4.974005718741877e-05,
|
| 510 |
+
"loss": 1.1082,
|
| 511 |
+
"step": 700
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 0.5465742879137798,
|
| 515 |
+
"grad_norm": 41.16048049926758,
|
| 516 |
+
"learning_rate": 4.972706004678971e-05,
|
| 517 |
+
"loss": 1.0006,
|
| 518 |
+
"step": 710
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.5542725173210161,
|
| 522 |
+
"grad_norm": 64.6050033569336,
|
| 523 |
+
"learning_rate": 4.9714062906160646e-05,
|
| 524 |
+
"loss": 1.2638,
|
| 525 |
+
"step": 720
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.5619707467282525,
|
| 529 |
+
"grad_norm": 35.458457946777344,
|
| 530 |
+
"learning_rate": 4.970106576553159e-05,
|
| 531 |
+
"loss": 1.0841,
|
| 532 |
+
"step": 730
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 0.5696689761354888,
|
| 536 |
+
"grad_norm": 33.32878494262695,
|
| 537 |
+
"learning_rate": 4.968806862490252e-05,
|
| 538 |
+
"loss": 1.2451,
|
| 539 |
+
"step": 740
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.5773672055427251,
|
| 543 |
+
"grad_norm": 26.29853630065918,
|
| 544 |
+
"learning_rate": 4.967507148427346e-05,
|
| 545 |
+
"loss": 1.0487,
|
| 546 |
+
"step": 750
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 0.5850654349499615,
|
| 550 |
+
"grad_norm": 6.7931084632873535,
|
| 551 |
+
"learning_rate": 4.96620743436444e-05,
|
| 552 |
+
"loss": 1.0836,
|
| 553 |
+
"step": 760
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 0.5927636643571979,
|
| 557 |
+
"grad_norm": 11.772964477539062,
|
| 558 |
+
"learning_rate": 4.9649077203015335e-05,
|
| 559 |
+
"loss": 0.9459,
|
| 560 |
+
"step": 770
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.6004618937644342,
|
| 564 |
+
"grad_norm": 83.65714263916016,
|
| 565 |
+
"learning_rate": 4.9636080062386276e-05,
|
| 566 |
+
"loss": 1.0921,
|
| 567 |
+
"step": 780
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.6081601231716706,
|
| 571 |
+
"grad_norm": 67.42829132080078,
|
| 572 |
+
"learning_rate": 4.962308292175721e-05,
|
| 573 |
+
"loss": 1.0981,
|
| 574 |
+
"step": 790
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 0.6158583525789069,
|
| 578 |
+
"grad_norm": 40.03876876831055,
|
| 579 |
+
"learning_rate": 4.961008578112815e-05,
|
| 580 |
+
"loss": 1.1155,
|
| 581 |
+
"step": 800
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.6235565819861432,
|
| 585 |
+
"grad_norm": 14.459259033203125,
|
| 586 |
+
"learning_rate": 4.9597088640499095e-05,
|
| 587 |
+
"loss": 1.2369,
|
| 588 |
+
"step": 810
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 0.6312548113933796,
|
| 592 |
+
"grad_norm": 43.21743392944336,
|
| 593 |
+
"learning_rate": 4.9584091499870037e-05,
|
| 594 |
+
"loss": 1.1154,
|
| 595 |
+
"step": 820
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 0.6389530408006159,
|
| 599 |
+
"grad_norm": 29.028535842895508,
|
| 600 |
+
"learning_rate": 4.957109435924097e-05,
|
| 601 |
+
"loss": 1.0684,
|
| 602 |
+
"step": 830
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.6466512702078522,
|
| 606 |
+
"grad_norm": 3421.910400390625,
|
| 607 |
+
"learning_rate": 4.9558097218611907e-05,
|
| 608 |
+
"loss": 1.1415,
|
| 609 |
+
"step": 840
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.6543494996150885,
|
| 613 |
+
"grad_norm": 50.08725357055664,
|
| 614 |
+
"learning_rate": 4.954510007798285e-05,
|
| 615 |
+
"loss": 1.5334,
|
| 616 |
+
"step": 850
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 0.6620477290223249,
|
| 620 |
+
"grad_norm": 31.24684715270996,
|
| 621 |
+
"learning_rate": 4.953210293735378e-05,
|
| 622 |
+
"loss": 1.1275,
|
| 623 |
+
"step": 860
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 0.6697459584295612,
|
| 627 |
+
"grad_norm": 56.685611724853516,
|
| 628 |
+
"learning_rate": 4.9519105796724725e-05,
|
| 629 |
+
"loss": 1.1105,
|
| 630 |
+
"step": 870
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 0.6774441878367975,
|
| 634 |
+
"grad_norm": 59.59196853637695,
|
| 635 |
+
"learning_rate": 4.950610865609566e-05,
|
| 636 |
+
"loss": 1.296,
|
| 637 |
+
"step": 880
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 0.6851424172440339,
|
| 641 |
+
"grad_norm": 59.091697692871094,
|
| 642 |
+
"learning_rate": 4.94931115154666e-05,
|
| 643 |
+
"loss": 1.0462,
|
| 644 |
+
"step": 890
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.6928406466512702,
|
| 648 |
+
"grad_norm": 25.288618087768555,
|
| 649 |
+
"learning_rate": 4.948011437483754e-05,
|
| 650 |
+
"loss": 0.9684,
|
| 651 |
+
"step": 900
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.7005388760585065,
|
| 655 |
+
"grad_norm": 34.281044006347656,
|
| 656 |
+
"learning_rate": 4.946711723420848e-05,
|
| 657 |
+
"loss": 1.0789,
|
| 658 |
+
"step": 910
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.7082371054657429,
|
| 662 |
+
"grad_norm": 12.016758918762207,
|
| 663 |
+
"learning_rate": 4.945412009357941e-05,
|
| 664 |
+
"loss": 1.1008,
|
| 665 |
+
"step": 920
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.7159353348729792,
|
| 669 |
+
"grad_norm": 23.659439086914062,
|
| 670 |
+
"learning_rate": 4.9441122952950355e-05,
|
| 671 |
+
"loss": 1.3602,
|
| 672 |
+
"step": 930
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 0.7236335642802155,
|
| 676 |
+
"grad_norm": 12.893980026245117,
|
| 677 |
+
"learning_rate": 4.942812581232129e-05,
|
| 678 |
+
"loss": 1.3067,
|
| 679 |
+
"step": 940
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 0.7313317936874519,
|
| 683 |
+
"grad_norm": 36.1422004699707,
|
| 684 |
+
"learning_rate": 4.941512867169223e-05,
|
| 685 |
+
"loss": 1.2795,
|
| 686 |
+
"step": 950
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.7390300230946882,
|
| 690 |
+
"grad_norm": 24.229007720947266,
|
| 691 |
+
"learning_rate": 4.940213153106317e-05,
|
| 692 |
+
"loss": 0.8846,
|
| 693 |
+
"step": 960
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.7467282525019245,
|
| 697 |
+
"grad_norm": 12.185916900634766,
|
| 698 |
+
"learning_rate": 4.938913439043411e-05,
|
| 699 |
+
"loss": 1.0618,
|
| 700 |
+
"step": 970
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.7544264819091608,
|
| 704 |
+
"grad_norm": 58.11173629760742,
|
| 705 |
+
"learning_rate": 4.9376137249805044e-05,
|
| 706 |
+
"loss": 1.0567,
|
| 707 |
+
"step": 980
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.7621247113163973,
|
| 711 |
+
"grad_norm": 17.494789123535156,
|
| 712 |
+
"learning_rate": 4.936314010917598e-05,
|
| 713 |
+
"loss": 0.9427,
|
| 714 |
+
"step": 990
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.7698229407236336,
|
| 718 |
+
"grad_norm": 21.812406539916992,
|
| 719 |
+
"learning_rate": 4.935014296854692e-05,
|
| 720 |
+
"loss": 1.0871,
|
| 721 |
+
"step": 1000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.7698229407236336,
|
| 725 |
+
"eval_accuracy": 0.5400307929176289,
|
| 726 |
+
"eval_f1": 0.5212757553318326,
|
| 727 |
+
"eval_loss": 1.0452667474746704,
|
| 728 |
+
"eval_precision": 0.5391802513141176,
|
| 729 |
+
"eval_recall": 0.5400307929176289,
|
| 730 |
+
"eval_runtime": 7.9557,
|
| 731 |
+
"eval_samples_per_second": 326.559,
|
| 732 |
+
"eval_steps_per_second": 40.851,
|
| 733 |
+
"step": 1000
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 0.7775211701308699,
|
| 737 |
+
"grad_norm": 68.46436309814453,
|
| 738 |
+
"learning_rate": 4.9337145827917855e-05,
|
| 739 |
+
"loss": 1.3711,
|
| 740 |
+
"step": 1010
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.7852193995381063,
|
| 744 |
+
"grad_norm": 25.05400848388672,
|
| 745 |
+
"learning_rate": 4.9324148687288804e-05,
|
| 746 |
+
"loss": 1.1931,
|
| 747 |
+
"step": 1020
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.7929176289453426,
|
| 751 |
+
"grad_norm": 24.60456085205078,
|
| 752 |
+
"learning_rate": 4.931115154665974e-05,
|
| 753 |
+
"loss": 1.1109,
|
| 754 |
+
"step": 1030
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.8006158583525789,
|
| 758 |
+
"grad_norm": 24.26364517211914,
|
| 759 |
+
"learning_rate": 4.929815440603068e-05,
|
| 760 |
+
"loss": 1.0804,
|
| 761 |
+
"step": 1040
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 0.8083140877598153,
|
| 765 |
+
"grad_norm": 22.994844436645508,
|
| 766 |
+
"learning_rate": 4.9285157265401615e-05,
|
| 767 |
+
"loss": 0.8045,
|
| 768 |
+
"step": 1050
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 0.8160123171670516,
|
| 772 |
+
"grad_norm": 16.76064682006836,
|
| 773 |
+
"learning_rate": 4.927216012477256e-05,
|
| 774 |
+
"loss": 1.1683,
|
| 775 |
+
"step": 1060
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 0.8237105465742879,
|
| 779 |
+
"grad_norm": 46.755393981933594,
|
| 780 |
+
"learning_rate": 4.925916298414349e-05,
|
| 781 |
+
"loss": 1.1455,
|
| 782 |
+
"step": 1070
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.8314087759815243,
|
| 786 |
+
"grad_norm": 46.636077880859375,
|
| 787 |
+
"learning_rate": 4.924616584351443e-05,
|
| 788 |
+
"loss": 0.9044,
|
| 789 |
+
"step": 1080
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.8391070053887606,
|
| 793 |
+
"grad_norm": 25.199445724487305,
|
| 794 |
+
"learning_rate": 4.923316870288537e-05,
|
| 795 |
+
"loss": 1.0345,
|
| 796 |
+
"step": 1090
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 0.8468052347959969,
|
| 800 |
+
"grad_norm": 37.11675262451172,
|
| 801 |
+
"learning_rate": 4.9220171562256304e-05,
|
| 802 |
+
"loss": 1.2491,
|
| 803 |
+
"step": 1100
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 0.8545034642032333,
|
| 807 |
+
"grad_norm": 19.960737228393555,
|
| 808 |
+
"learning_rate": 4.9207174421627246e-05,
|
| 809 |
+
"loss": 0.9029,
|
| 810 |
+
"step": 1110
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.8622016936104696,
|
| 814 |
+
"grad_norm": 81.35137939453125,
|
| 815 |
+
"learning_rate": 4.919417728099818e-05,
|
| 816 |
+
"loss": 1.2947,
|
| 817 |
+
"step": 1120
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 0.8698999230177059,
|
| 821 |
+
"grad_norm": 32.889427185058594,
|
| 822 |
+
"learning_rate": 4.918118014036912e-05,
|
| 823 |
+
"loss": 1.1014,
|
| 824 |
+
"step": 1130
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.8775981524249422,
|
| 828 |
+
"grad_norm": 35.65376281738281,
|
| 829 |
+
"learning_rate": 4.916818299974006e-05,
|
| 830 |
+
"loss": 1.1406,
|
| 831 |
+
"step": 1140
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.8852963818321786,
|
| 835 |
+
"grad_norm": 309.11114501953125,
|
| 836 |
+
"learning_rate": 4.9155185859111e-05,
|
| 837 |
+
"loss": 0.9684,
|
| 838 |
+
"step": 1150
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 0.8929946112394149,
|
| 842 |
+
"grad_norm": 19.624103546142578,
|
| 843 |
+
"learning_rate": 4.9142188718481934e-05,
|
| 844 |
+
"loss": 0.9999,
|
| 845 |
+
"step": 1160
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 0.9006928406466512,
|
| 849 |
+
"grad_norm": 35.4536018371582,
|
| 850 |
+
"learning_rate": 4.9129191577852876e-05,
|
| 851 |
+
"loss": 1.3579,
|
| 852 |
+
"step": 1170
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 0.9083910700538876,
|
| 856 |
+
"grad_norm": 94.22552490234375,
|
| 857 |
+
"learning_rate": 4.911619443722381e-05,
|
| 858 |
+
"loss": 1.0926,
|
| 859 |
+
"step": 1180
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 0.9160892994611239,
|
| 863 |
+
"grad_norm": 23.519554138183594,
|
| 864 |
+
"learning_rate": 4.910319729659475e-05,
|
| 865 |
+
"loss": 1.1782,
|
| 866 |
+
"step": 1190
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.9237875288683602,
|
| 870 |
+
"grad_norm": 25.848222732543945,
|
| 871 |
+
"learning_rate": 4.909020015596569e-05,
|
| 872 |
+
"loss": 1.0085,
|
| 873 |
+
"step": 1200
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.9314857582755967,
|
| 877 |
+
"grad_norm": 40.45154571533203,
|
| 878 |
+
"learning_rate": 4.907720301533663e-05,
|
| 879 |
+
"loss": 1.0384,
|
| 880 |
+
"step": 1210
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 0.939183987682833,
|
| 884 |
+
"grad_norm": 26.536237716674805,
|
| 885 |
+
"learning_rate": 4.9064205874707564e-05,
|
| 886 |
+
"loss": 1.006,
|
| 887 |
+
"step": 1220
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 0.9468822170900693,
|
| 891 |
+
"grad_norm": 17.983915328979492,
|
| 892 |
+
"learning_rate": 4.9051208734078506e-05,
|
| 893 |
+
"loss": 1.1031,
|
| 894 |
+
"step": 1230
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.9545804464973057,
|
| 898 |
+
"grad_norm": 46.20643615722656,
|
| 899 |
+
"learning_rate": 4.903821159344945e-05,
|
| 900 |
+
"loss": 0.9322,
|
| 901 |
+
"step": 1240
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.962278675904542,
|
| 905 |
+
"grad_norm": 59.26303482055664,
|
| 906 |
+
"learning_rate": 4.902521445282038e-05,
|
| 907 |
+
"loss": 1.1146,
|
| 908 |
+
"step": 1250
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.9699769053117783,
|
| 912 |
+
"grad_norm": 42.50138473510742,
|
| 913 |
+
"learning_rate": 4.9012217312191324e-05,
|
| 914 |
+
"loss": 1.1769,
|
| 915 |
+
"step": 1260
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.9776751347190146,
|
| 919 |
+
"grad_norm": 44.46087646484375,
|
| 920 |
+
"learning_rate": 4.899922017156226e-05,
|
| 921 |
+
"loss": 1.113,
|
| 922 |
+
"step": 1270
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.985373364126251,
|
| 926 |
+
"grad_norm": 24.701412200927734,
|
| 927 |
+
"learning_rate": 4.89862230309332e-05,
|
| 928 |
+
"loss": 1.0445,
|
| 929 |
+
"step": 1280
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.9930715935334873,
|
| 933 |
+
"grad_norm": 19.296676635742188,
|
| 934 |
+
"learning_rate": 4.8973225890304136e-05,
|
| 935 |
+
"loss": 1.075,
|
| 936 |
+
"step": 1290
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 1.0007698229407236,
|
| 940 |
+
"grad_norm": 24.70547866821289,
|
| 941 |
+
"learning_rate": 4.896022874967508e-05,
|
| 942 |
+
"loss": 0.8709,
|
| 943 |
+
"step": 1300
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.0084680523479599,
|
| 947 |
+
"grad_norm": 35.57961654663086,
|
| 948 |
+
"learning_rate": 4.894723160904601e-05,
|
| 949 |
+
"loss": 1.0378,
|
| 950 |
+
"step": 1310
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 1.0161662817551964,
|
| 954 |
+
"grad_norm": 30.505125045776367,
|
| 955 |
+
"learning_rate": 4.893423446841695e-05,
|
| 956 |
+
"loss": 0.992,
|
| 957 |
+
"step": 1320
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 1.0238645111624327,
|
| 961 |
+
"grad_norm": 58.703643798828125,
|
| 962 |
+
"learning_rate": 4.892123732778789e-05,
|
| 963 |
+
"loss": 1.1885,
|
| 964 |
+
"step": 1330
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 1.031562740569669,
|
| 968 |
+
"grad_norm": 28.3648681640625,
|
| 969 |
+
"learning_rate": 4.8908240187158824e-05,
|
| 970 |
+
"loss": 1.2337,
|
| 971 |
+
"step": 1340
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 1.0392609699769053,
|
| 975 |
+
"grad_norm": 24.309017181396484,
|
| 976 |
+
"learning_rate": 4.8895243046529766e-05,
|
| 977 |
+
"loss": 1.0578,
|
| 978 |
+
"step": 1350
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 1.0469591993841416,
|
| 982 |
+
"grad_norm": 26.426084518432617,
|
| 983 |
+
"learning_rate": 4.88822459059007e-05,
|
| 984 |
+
"loss": 0.9974,
|
| 985 |
+
"step": 1360
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.054657428791378,
|
| 989 |
+
"grad_norm": 27.277679443359375,
|
| 990 |
+
"learning_rate": 4.886924876527164e-05,
|
| 991 |
+
"loss": 1.0032,
|
| 992 |
+
"step": 1370
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 1.0623556581986142,
|
| 996 |
+
"grad_norm": 22.838211059570312,
|
| 997 |
+
"learning_rate": 4.885625162464258e-05,
|
| 998 |
+
"loss": 1.1148,
|
| 999 |
+
"step": 1380
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 1.0700538876058507,
|
| 1003 |
+
"grad_norm": 41.72176742553711,
|
| 1004 |
+
"learning_rate": 4.884325448401352e-05,
|
| 1005 |
+
"loss": 0.9433,
|
| 1006 |
+
"step": 1390
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 1.077752117013087,
|
| 1010 |
+
"grad_norm": 72.71926879882812,
|
| 1011 |
+
"learning_rate": 4.8830257343384455e-05,
|
| 1012 |
+
"loss": 0.8138,
|
| 1013 |
+
"step": 1400
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.0854503464203233,
|
| 1017 |
+
"grad_norm": 20.042091369628906,
|
| 1018 |
+
"learning_rate": 4.8817260202755396e-05,
|
| 1019 |
+
"loss": 1.0195,
|
| 1020 |
+
"step": 1410
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 1.0931485758275596,
|
| 1024 |
+
"grad_norm": 20.96778678894043,
|
| 1025 |
+
"learning_rate": 4.880426306212633e-05,
|
| 1026 |
+
"loss": 1.0269,
|
| 1027 |
+
"step": 1420
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.100846805234796,
|
| 1031 |
+
"grad_norm": 32.662322998046875,
|
| 1032 |
+
"learning_rate": 4.879126592149727e-05,
|
| 1033 |
+
"loss": 0.891,
|
| 1034 |
+
"step": 1430
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 1.1085450346420322,
|
| 1038 |
+
"grad_norm": 19.693790435791016,
|
| 1039 |
+
"learning_rate": 4.8778268780868215e-05,
|
| 1040 |
+
"loss": 1.054,
|
| 1041 |
+
"step": 1440
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 1.1162432640492688,
|
| 1045 |
+
"grad_norm": 20.91014289855957,
|
| 1046 |
+
"learning_rate": 4.876527164023915e-05,
|
| 1047 |
+
"loss": 0.9135,
|
| 1048 |
+
"step": 1450
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 1.123941493456505,
|
| 1052 |
+
"grad_norm": 52.75801086425781,
|
| 1053 |
+
"learning_rate": 4.875227449961009e-05,
|
| 1054 |
+
"loss": 1.0464,
|
| 1055 |
+
"step": 1460
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 1.1316397228637414,
|
| 1059 |
+
"grad_norm": 42.71292495727539,
|
| 1060 |
+
"learning_rate": 4.8739277358981026e-05,
|
| 1061 |
+
"loss": 1.0075,
|
| 1062 |
+
"step": 1470
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 1.1393379522709777,
|
| 1066 |
+
"grad_norm": 21.153303146362305,
|
| 1067 |
+
"learning_rate": 4.872628021835197e-05,
|
| 1068 |
+
"loss": 1.2111,
|
| 1069 |
+
"step": 1480
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.147036181678214,
|
| 1073 |
+
"grad_norm": 44.54867172241211,
|
| 1074 |
+
"learning_rate": 4.87132830777229e-05,
|
| 1075 |
+
"loss": 0.8313,
|
| 1076 |
+
"step": 1490
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 1.1547344110854503,
|
| 1080 |
+
"grad_norm": 48.92774963378906,
|
| 1081 |
+
"learning_rate": 4.8700285937093845e-05,
|
| 1082 |
+
"loss": 1.0748,
|
| 1083 |
+
"step": 1500
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 1.1547344110854503,
|
| 1087 |
+
"eval_accuracy": 0.5658198614318707,
|
| 1088 |
+
"eval_f1": 0.5450550292077301,
|
| 1089 |
+
"eval_loss": 0.9736452102661133,
|
| 1090 |
+
"eval_precision": 0.5647189867888521,
|
| 1091 |
+
"eval_recall": 0.5658198614318707,
|
| 1092 |
+
"eval_runtime": 8.0094,
|
| 1093 |
+
"eval_samples_per_second": 324.367,
|
| 1094 |
+
"eval_steps_per_second": 40.577,
|
| 1095 |
+
"step": 1500
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.1624326404926868,
|
| 1099 |
+
"grad_norm": 39.90788650512695,
|
| 1100 |
+
"learning_rate": 4.868728879646478e-05,
|
| 1101 |
+
"loss": 1.0338,
|
| 1102 |
+
"step": 1510
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.170130869899923,
|
| 1106 |
+
"grad_norm": 35.67189025878906,
|
| 1107 |
+
"learning_rate": 4.867429165583572e-05,
|
| 1108 |
+
"loss": 1.0185,
|
| 1109 |
+
"step": 1520
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.1778290993071594,
|
| 1113 |
+
"grad_norm": 76.6269302368164,
|
| 1114 |
+
"learning_rate": 4.8661294515206657e-05,
|
| 1115 |
+
"loss": 1.0013,
|
| 1116 |
+
"step": 1530
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.1855273287143957,
|
| 1120 |
+
"grad_norm": 166.01284790039062,
|
| 1121 |
+
"learning_rate": 4.86482973745776e-05,
|
| 1122 |
+
"loss": 1.0846,
|
| 1123 |
+
"step": 1540
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.193225558121632,
|
| 1127 |
+
"grad_norm": 26.522552490234375,
|
| 1128 |
+
"learning_rate": 4.863530023394853e-05,
|
| 1129 |
+
"loss": 1.012,
|
| 1130 |
+
"step": 1550
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.2009237875288683,
|
| 1134 |
+
"grad_norm": 23.47018051147461,
|
| 1135 |
+
"learning_rate": 4.862230309331947e-05,
|
| 1136 |
+
"loss": 0.9853,
|
| 1137 |
+
"step": 1560
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.2086220169361046,
|
| 1141 |
+
"grad_norm": 16.074838638305664,
|
| 1142 |
+
"learning_rate": 4.860930595269041e-05,
|
| 1143 |
+
"loss": 0.9196,
|
| 1144 |
+
"step": 1570
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.2163202463433411,
|
| 1148 |
+
"grad_norm": 21.358341217041016,
|
| 1149 |
+
"learning_rate": 4.8596308812061345e-05,
|
| 1150 |
+
"loss": 1.0509,
|
| 1151 |
+
"step": 1580
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.2240184757505774,
|
| 1155 |
+
"grad_norm": 28.035846710205078,
|
| 1156 |
+
"learning_rate": 4.858331167143229e-05,
|
| 1157 |
+
"loss": 0.9299,
|
| 1158 |
+
"step": 1590
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.2317167051578137,
|
| 1162 |
+
"grad_norm": 23.313568115234375,
|
| 1163 |
+
"learning_rate": 4.857031453080322e-05,
|
| 1164 |
+
"loss": 0.9581,
|
| 1165 |
+
"step": 1600
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.23941493456505,
|
| 1169 |
+
"grad_norm": 26.382869720458984,
|
| 1170 |
+
"learning_rate": 4.8557317390174163e-05,
|
| 1171 |
+
"loss": 0.8374,
|
| 1172 |
+
"step": 1610
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.2471131639722863,
|
| 1176 |
+
"grad_norm": 62.66410446166992,
|
| 1177 |
+
"learning_rate": 4.85443202495451e-05,
|
| 1178 |
+
"loss": 1.1268,
|
| 1179 |
+
"step": 1620
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.2548113933795226,
|
| 1183 |
+
"grad_norm": 32.7457160949707,
|
| 1184 |
+
"learning_rate": 4.853132310891604e-05,
|
| 1185 |
+
"loss": 0.7796,
|
| 1186 |
+
"step": 1630
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.2625096227867592,
|
| 1190 |
+
"grad_norm": 45.75529861450195,
|
| 1191 |
+
"learning_rate": 4.851832596828698e-05,
|
| 1192 |
+
"loss": 1.0265,
|
| 1193 |
+
"step": 1640
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.2702078521939955,
|
| 1197 |
+
"grad_norm": 73.32510375976562,
|
| 1198 |
+
"learning_rate": 4.850532882765792e-05,
|
| 1199 |
+
"loss": 1.3302,
|
| 1200 |
+
"step": 1650
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.2779060816012318,
|
| 1204 |
+
"grad_norm": 51.92557144165039,
|
| 1205 |
+
"learning_rate": 4.849233168702886e-05,
|
| 1206 |
+
"loss": 1.2056,
|
| 1207 |
+
"step": 1660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.285604311008468,
|
| 1211 |
+
"grad_norm": 9.508386611938477,
|
| 1212 |
+
"learning_rate": 4.8479334546399794e-05,
|
| 1213 |
+
"loss": 0.8817,
|
| 1214 |
+
"step": 1670
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.2933025404157044,
|
| 1218 |
+
"grad_norm": 19.91646957397461,
|
| 1219 |
+
"learning_rate": 4.8466337405770735e-05,
|
| 1220 |
+
"loss": 1.1081,
|
| 1221 |
+
"step": 1680
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.3010007698229407,
|
| 1225 |
+
"grad_norm": 22.991626739501953,
|
| 1226 |
+
"learning_rate": 4.845334026514167e-05,
|
| 1227 |
+
"loss": 0.9953,
|
| 1228 |
+
"step": 1690
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.308698999230177,
|
| 1232 |
+
"grad_norm": 36.05852127075195,
|
| 1233 |
+
"learning_rate": 4.844034312451261e-05,
|
| 1234 |
+
"loss": 0.9753,
|
| 1235 |
+
"step": 1700
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.3163972286374133,
|
| 1239 |
+
"grad_norm": 30.5816593170166,
|
| 1240 |
+
"learning_rate": 4.842734598388355e-05,
|
| 1241 |
+
"loss": 1.0043,
|
| 1242 |
+
"step": 1710
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.3240954580446498,
|
| 1246 |
+
"grad_norm": 32.22935485839844,
|
| 1247 |
+
"learning_rate": 4.841434884325449e-05,
|
| 1248 |
+
"loss": 0.977,
|
| 1249 |
+
"step": 1720
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.331793687451886,
|
| 1253 |
+
"grad_norm": 93.55306243896484,
|
| 1254 |
+
"learning_rate": 4.8401351702625424e-05,
|
| 1255 |
+
"loss": 0.959,
|
| 1256 |
+
"step": 1730
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.3394919168591224,
|
| 1260 |
+
"grad_norm": 20.172109603881836,
|
| 1261 |
+
"learning_rate": 4.8388354561996365e-05,
|
| 1262 |
+
"loss": 0.8522,
|
| 1263 |
+
"step": 1740
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.3471901462663587,
|
| 1267 |
+
"grad_norm": 34.432254791259766,
|
| 1268 |
+
"learning_rate": 4.83753574213673e-05,
|
| 1269 |
+
"loss": 0.9968,
|
| 1270 |
+
"step": 1750
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.354888375673595,
|
| 1274 |
+
"grad_norm": 9.962430000305176,
|
| 1275 |
+
"learning_rate": 4.836236028073824e-05,
|
| 1276 |
+
"loss": 0.8846,
|
| 1277 |
+
"step": 1760
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.3625866050808315,
|
| 1281 |
+
"grad_norm": 16.258909225463867,
|
| 1282 |
+
"learning_rate": 4.834936314010918e-05,
|
| 1283 |
+
"loss": 0.9685,
|
| 1284 |
+
"step": 1770
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.3702848344880678,
|
| 1288 |
+
"grad_norm": 600.8819580078125,
|
| 1289 |
+
"learning_rate": 4.833636599948012e-05,
|
| 1290 |
+
"loss": 1.2891,
|
| 1291 |
+
"step": 1780
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.3779830638953041,
|
| 1295 |
+
"grad_norm": 19.667699813842773,
|
| 1296 |
+
"learning_rate": 4.8323368858851054e-05,
|
| 1297 |
+
"loss": 0.9624,
|
| 1298 |
+
"step": 1790
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.3856812933025404,
|
| 1302 |
+
"grad_norm": 28.33806037902832,
|
| 1303 |
+
"learning_rate": 4.831037171822199e-05,
|
| 1304 |
+
"loss": 0.9677,
|
| 1305 |
+
"step": 1800
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.3933795227097767,
|
| 1309 |
+
"grad_norm": 17.19879722595215,
|
| 1310 |
+
"learning_rate": 4.829737457759293e-05,
|
| 1311 |
+
"loss": 0.7864,
|
| 1312 |
+
"step": 1810
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.401077752117013,
|
| 1316 |
+
"grad_norm": 68.86786651611328,
|
| 1317 |
+
"learning_rate": 4.8284377436963865e-05,
|
| 1318 |
+
"loss": 1.0756,
|
| 1319 |
+
"step": 1820
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.4087759815242493,
|
| 1323 |
+
"grad_norm": 19.253087997436523,
|
| 1324 |
+
"learning_rate": 4.827138029633481e-05,
|
| 1325 |
+
"loss": 0.902,
|
| 1326 |
+
"step": 1830
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.4164742109314856,
|
| 1330 |
+
"grad_norm": 22.85369300842285,
|
| 1331 |
+
"learning_rate": 4.825838315570575e-05,
|
| 1332 |
+
"loss": 0.9561,
|
| 1333 |
+
"step": 1840
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.4241724403387221,
|
| 1337 |
+
"grad_norm": 34.35043716430664,
|
| 1338 |
+
"learning_rate": 4.824538601507669e-05,
|
| 1339 |
+
"loss": 0.938,
|
| 1340 |
+
"step": 1850
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.4318706697459584,
|
| 1344 |
+
"grad_norm": 29.028274536132812,
|
| 1345 |
+
"learning_rate": 4.8232388874447626e-05,
|
| 1346 |
+
"loss": 0.9659,
|
| 1347 |
+
"step": 1860
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.4395688991531947,
|
| 1351 |
+
"grad_norm": 15.733872413635254,
|
| 1352 |
+
"learning_rate": 4.821939173381857e-05,
|
| 1353 |
+
"loss": 0.955,
|
| 1354 |
+
"step": 1870
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.447267128560431,
|
| 1358 |
+
"grad_norm": 45.1998176574707,
|
| 1359 |
+
"learning_rate": 4.82063945931895e-05,
|
| 1360 |
+
"loss": 1.0997,
|
| 1361 |
+
"step": 1880
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.4549653579676676,
|
| 1365 |
+
"grad_norm": 62.95772171020508,
|
| 1366 |
+
"learning_rate": 4.819339745256044e-05,
|
| 1367 |
+
"loss": 0.9975,
|
| 1368 |
+
"step": 1890
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.4626635873749039,
|
| 1372 |
+
"grad_norm": 13.520893096923828,
|
| 1373 |
+
"learning_rate": 4.818040031193138e-05,
|
| 1374 |
+
"loss": 0.9224,
|
| 1375 |
+
"step": 1900
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.4703618167821402,
|
| 1379 |
+
"grad_norm": 44.204837799072266,
|
| 1380 |
+
"learning_rate": 4.8167403171302314e-05,
|
| 1381 |
+
"loss": 1.0823,
|
| 1382 |
+
"step": 1910
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.4780600461893765,
|
| 1386 |
+
"grad_norm": 18.848365783691406,
|
| 1387 |
+
"learning_rate": 4.8154406030673256e-05,
|
| 1388 |
+
"loss": 0.8767,
|
| 1389 |
+
"step": 1920
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.4857582755966128,
|
| 1393 |
+
"grad_norm": 48.03689193725586,
|
| 1394 |
+
"learning_rate": 4.814140889004419e-05,
|
| 1395 |
+
"loss": 1.1274,
|
| 1396 |
+
"step": 1930
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.493456505003849,
|
| 1400 |
+
"grad_norm": 29.64519691467285,
|
| 1401 |
+
"learning_rate": 4.812841174941513e-05,
|
| 1402 |
+
"loss": 0.8109,
|
| 1403 |
+
"step": 1940
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.5011547344110854,
|
| 1407 |
+
"grad_norm": 56.1988639831543,
|
| 1408 |
+
"learning_rate": 4.811541460878607e-05,
|
| 1409 |
+
"loss": 1.1295,
|
| 1410 |
+
"step": 1950
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.5088529638183217,
|
| 1414 |
+
"grad_norm": 39.60947036743164,
|
| 1415 |
+
"learning_rate": 4.810241746815701e-05,
|
| 1416 |
+
"loss": 1.1057,
|
| 1417 |
+
"step": 1960
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.516551193225558,
|
| 1421 |
+
"grad_norm": 14.454174041748047,
|
| 1422 |
+
"learning_rate": 4.8089420327527944e-05,
|
| 1423 |
+
"loss": 1.098,
|
| 1424 |
+
"step": 1970
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.5242494226327945,
|
| 1428 |
+
"grad_norm": 18.127788543701172,
|
| 1429 |
+
"learning_rate": 4.8076423186898886e-05,
|
| 1430 |
+
"loss": 0.947,
|
| 1431 |
+
"step": 1980
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.5319476520400308,
|
| 1435 |
+
"grad_norm": 30.042423248291016,
|
| 1436 |
+
"learning_rate": 4.806342604626982e-05,
|
| 1437 |
+
"loss": 0.8846,
|
| 1438 |
+
"step": 1990
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.539645881447267,
|
| 1442 |
+
"grad_norm": 31.532861709594727,
|
| 1443 |
+
"learning_rate": 4.805042890564076e-05,
|
| 1444 |
+
"loss": 0.908,
|
| 1445 |
+
"step": 2000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.539645881447267,
|
| 1449 |
+
"eval_accuracy": 0.5873749037721324,
|
| 1450 |
+
"eval_f1": 0.5698601671652778,
|
| 1451 |
+
"eval_loss": 0.9315417408943176,
|
| 1452 |
+
"eval_precision": 0.5808478011880224,
|
| 1453 |
+
"eval_recall": 0.5873749037721324,
|
| 1454 |
+
"eval_runtime": 8.0056,
|
| 1455 |
+
"eval_samples_per_second": 324.524,
|
| 1456 |
+
"eval_steps_per_second": 40.597,
|
| 1457 |
+
"step": 2000
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 1.5473441108545036,
|
| 1461 |
+
"grad_norm": 87.31982421875,
|
| 1462 |
+
"learning_rate": 4.80374317650117e-05,
|
| 1463 |
+
"loss": 1.2135,
|
| 1464 |
+
"step": 2010
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 1.55504234026174,
|
| 1468 |
+
"grad_norm": 36.29372787475586,
|
| 1469 |
+
"learning_rate": 4.802443462438263e-05,
|
| 1470 |
+
"loss": 0.9446,
|
| 1471 |
+
"step": 2020
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.5627405696689762,
|
| 1475 |
+
"grad_norm": 65.0682144165039,
|
| 1476 |
+
"learning_rate": 4.8011437483753574e-05,
|
| 1477 |
+
"loss": 0.942,
|
| 1478 |
+
"step": 2030
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.5704387990762125,
|
| 1482 |
+
"grad_norm": 7.867269992828369,
|
| 1483 |
+
"learning_rate": 4.799844034312451e-05,
|
| 1484 |
+
"loss": 0.9343,
|
| 1485 |
+
"step": 2040
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 1.5781370284834488,
|
| 1489 |
+
"grad_norm": 18.922204971313477,
|
| 1490 |
+
"learning_rate": 4.798544320249546e-05,
|
| 1491 |
+
"loss": 1.0242,
|
| 1492 |
+
"step": 2050
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.5858352578906851,
|
| 1496 |
+
"grad_norm": 22.182636260986328,
|
| 1497 |
+
"learning_rate": 4.797244606186639e-05,
|
| 1498 |
+
"loss": 0.8997,
|
| 1499 |
+
"step": 2060
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 1.5935334872979214,
|
| 1503 |
+
"grad_norm": 13.871538162231445,
|
| 1504 |
+
"learning_rate": 4.7959448921237335e-05,
|
| 1505 |
+
"loss": 1.0238,
|
| 1506 |
+
"step": 2070
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 1.6012317167051577,
|
| 1510 |
+
"grad_norm": 24.956518173217773,
|
| 1511 |
+
"learning_rate": 4.794645178060827e-05,
|
| 1512 |
+
"loss": 0.8555,
|
| 1513 |
+
"step": 2080
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.608929946112394,
|
| 1517 |
+
"grad_norm": 27.833717346191406,
|
| 1518 |
+
"learning_rate": 4.793345463997921e-05,
|
| 1519 |
+
"loss": 1.1057,
|
| 1520 |
+
"step": 2090
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.6166281755196303,
|
| 1524 |
+
"grad_norm": 15.595669746398926,
|
| 1525 |
+
"learning_rate": 4.7920457499350146e-05,
|
| 1526 |
+
"loss": 0.8761,
|
| 1527 |
+
"step": 2100
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.6243264049268669,
|
| 1531 |
+
"grad_norm": 33.80525207519531,
|
| 1532 |
+
"learning_rate": 4.790746035872109e-05,
|
| 1533 |
+
"loss": 0.9889,
|
| 1534 |
+
"step": 2110
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.6320246343341032,
|
| 1538 |
+
"grad_norm": 21.781145095825195,
|
| 1539 |
+
"learning_rate": 4.789446321809202e-05,
|
| 1540 |
+
"loss": 1.055,
|
| 1541 |
+
"step": 2120
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.6397228637413395,
|
| 1545 |
+
"grad_norm": 29.12350082397461,
|
| 1546 |
+
"learning_rate": 4.788146607746296e-05,
|
| 1547 |
+
"loss": 1.0033,
|
| 1548 |
+
"step": 2130
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 1.647421093148576,
|
| 1552 |
+
"grad_norm": 20.39841079711914,
|
| 1553 |
+
"learning_rate": 4.78684689368339e-05,
|
| 1554 |
+
"loss": 1.0698,
|
| 1555 |
+
"step": 2140
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.6551193225558123,
|
| 1559 |
+
"grad_norm": 9.309553146362305,
|
| 1560 |
+
"learning_rate": 4.7855471796204835e-05,
|
| 1561 |
+
"loss": 0.9092,
|
| 1562 |
+
"step": 2150
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 1.6628175519630486,
|
| 1566 |
+
"grad_norm": 28.19939613342285,
|
| 1567 |
+
"learning_rate": 4.7842474655575776e-05,
|
| 1568 |
+
"loss": 0.9727,
|
| 1569 |
+
"step": 2160
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.670515781370285,
|
| 1573 |
+
"grad_norm": 18.627004623413086,
|
| 1574 |
+
"learning_rate": 4.782947751494671e-05,
|
| 1575 |
+
"loss": 0.8224,
|
| 1576 |
+
"step": 2170
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.6782140107775212,
|
| 1580 |
+
"grad_norm": 15.454500198364258,
|
| 1581 |
+
"learning_rate": 4.781648037431765e-05,
|
| 1582 |
+
"loss": 0.9459,
|
| 1583 |
+
"step": 2180
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.6859122401847575,
|
| 1587 |
+
"grad_norm": 25.361587524414062,
|
| 1588 |
+
"learning_rate": 4.780348323368859e-05,
|
| 1589 |
+
"loss": 0.8389,
|
| 1590 |
+
"step": 2190
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 1.6936104695919938,
|
| 1594 |
+
"grad_norm": 48.25446701049805,
|
| 1595 |
+
"learning_rate": 4.779048609305953e-05,
|
| 1596 |
+
"loss": 1.1498,
|
| 1597 |
+
"step": 2200
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.70130869899923,
|
| 1601 |
+
"grad_norm": 45.576908111572266,
|
| 1602 |
+
"learning_rate": 4.7777488952430465e-05,
|
| 1603 |
+
"loss": 0.7871,
|
| 1604 |
+
"step": 2210
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 1.7090069284064664,
|
| 1608 |
+
"grad_norm": 28.02678871154785,
|
| 1609 |
+
"learning_rate": 4.7764491811801407e-05,
|
| 1610 |
+
"loss": 0.8891,
|
| 1611 |
+
"step": 2220
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.7167051578137027,
|
| 1615 |
+
"grad_norm": 17.72388458251953,
|
| 1616 |
+
"learning_rate": 4.775149467117234e-05,
|
| 1617 |
+
"loss": 1.0627,
|
| 1618 |
+
"step": 2230
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.7244033872209392,
|
| 1622 |
+
"grad_norm": 14.796937942504883,
|
| 1623 |
+
"learning_rate": 4.773849753054328e-05,
|
| 1624 |
+
"loss": 0.8403,
|
| 1625 |
+
"step": 2240
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.7321016166281755,
|
| 1629 |
+
"grad_norm": 18.26534652709961,
|
| 1630 |
+
"learning_rate": 4.7725500389914225e-05,
|
| 1631 |
+
"loss": 1.0399,
|
| 1632 |
+
"step": 2250
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 1.7397998460354118,
|
| 1636 |
+
"grad_norm": 12.606345176696777,
|
| 1637 |
+
"learning_rate": 4.771250324928516e-05,
|
| 1638 |
+
"loss": 0.7362,
|
| 1639 |
+
"step": 2260
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.7474980754426483,
|
| 1643 |
+
"grad_norm": 10.439709663391113,
|
| 1644 |
+
"learning_rate": 4.76995061086561e-05,
|
| 1645 |
+
"loss": 1.0418,
|
| 1646 |
+
"step": 2270
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 1.7551963048498846,
|
| 1650 |
+
"grad_norm": 24.162647247314453,
|
| 1651 |
+
"learning_rate": 4.768650896802704e-05,
|
| 1652 |
+
"loss": 0.9489,
|
| 1653 |
+
"step": 2280
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.762894534257121,
|
| 1657 |
+
"grad_norm": 34.678993225097656,
|
| 1658 |
+
"learning_rate": 4.767351182739798e-05,
|
| 1659 |
+
"loss": 0.8907,
|
| 1660 |
+
"step": 2290
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.7705927636643572,
|
| 1664 |
+
"grad_norm": 22.55309295654297,
|
| 1665 |
+
"learning_rate": 4.7660514686768913e-05,
|
| 1666 |
+
"loss": 0.8239,
|
| 1667 |
+
"step": 2300
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.7782909930715936,
|
| 1671 |
+
"grad_norm": 22.526445388793945,
|
| 1672 |
+
"learning_rate": 4.7647517546139855e-05,
|
| 1673 |
+
"loss": 1.0999,
|
| 1674 |
+
"step": 2310
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 1.7859892224788299,
|
| 1678 |
+
"grad_norm": 19.45115089416504,
|
| 1679 |
+
"learning_rate": 4.763452040551079e-05,
|
| 1680 |
+
"loss": 0.8858,
|
| 1681 |
+
"step": 2320
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.7936874518860662,
|
| 1685 |
+
"grad_norm": 29.1611270904541,
|
| 1686 |
+
"learning_rate": 4.762152326488173e-05,
|
| 1687 |
+
"loss": 1.0418,
|
| 1688 |
+
"step": 2330
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 1.8013856812933025,
|
| 1692 |
+
"grad_norm": 22.248546600341797,
|
| 1693 |
+
"learning_rate": 4.760852612425267e-05,
|
| 1694 |
+
"loss": 0.8722,
|
| 1695 |
+
"step": 2340
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 1.8090839107005388,
|
| 1699 |
+
"grad_norm": 78.22930908203125,
|
| 1700 |
+
"learning_rate": 4.75955289836236e-05,
|
| 1701 |
+
"loss": 0.8606,
|
| 1702 |
+
"step": 2350
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 1.816782140107775,
|
| 1706 |
+
"grad_norm": 16.559967041015625,
|
| 1707 |
+
"learning_rate": 4.7582531842994544e-05,
|
| 1708 |
+
"loss": 0.9108,
|
| 1709 |
+
"step": 2360
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.8244803695150116,
|
| 1713 |
+
"grad_norm": 20.415922164916992,
|
| 1714 |
+
"learning_rate": 4.756953470236548e-05,
|
| 1715 |
+
"loss": 1.0512,
|
| 1716 |
+
"step": 2370
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 1.8321785989222479,
|
| 1720 |
+
"grad_norm": 34.354740142822266,
|
| 1721 |
+
"learning_rate": 4.755653756173642e-05,
|
| 1722 |
+
"loss": 0.9482,
|
| 1723 |
+
"step": 2380
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 1.8398768283294842,
|
| 1727 |
+
"grad_norm": 26.14980697631836,
|
| 1728 |
+
"learning_rate": 4.7543540421107355e-05,
|
| 1729 |
+
"loss": 0.9213,
|
| 1730 |
+
"step": 2390
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 1.8475750577367207,
|
| 1734 |
+
"grad_norm": 137.22198486328125,
|
| 1735 |
+
"learning_rate": 4.75305432804783e-05,
|
| 1736 |
+
"loss": 1.0509,
|
| 1737 |
+
"step": 2400
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 1.855273287143957,
|
| 1741 |
+
"grad_norm": 82.37857818603516,
|
| 1742 |
+
"learning_rate": 4.751754613984923e-05,
|
| 1743 |
+
"loss": 0.8466,
|
| 1744 |
+
"step": 2410
|
| 1745 |
+
},
|
| 1746 |
+
{
|
| 1747 |
+
"epoch": 1.8629715165511933,
|
| 1748 |
+
"grad_norm": 31.272960662841797,
|
| 1749 |
+
"learning_rate": 4.7504548999220174e-05,
|
| 1750 |
+
"loss": 0.9717,
|
| 1751 |
+
"step": 2420
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.8706697459584296,
|
| 1755 |
+
"grad_norm": 33.75693893432617,
|
| 1756 |
+
"learning_rate": 4.749155185859111e-05,
|
| 1757 |
+
"loss": 1.0565,
|
| 1758 |
+
"step": 2430
|
| 1759 |
+
},
|
| 1760 |
+
{
|
| 1761 |
+
"epoch": 1.878367975365666,
|
| 1762 |
+
"grad_norm": 21.18593406677246,
|
| 1763 |
+
"learning_rate": 4.747855471796205e-05,
|
| 1764 |
+
"loss": 0.8595,
|
| 1765 |
+
"step": 2440
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 1.8860662047729022,
|
| 1769 |
+
"grad_norm": 23.71518898010254,
|
| 1770 |
+
"learning_rate": 4.7465557577332985e-05,
|
| 1771 |
+
"loss": 0.8621,
|
| 1772 |
+
"step": 2450
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 1.8937644341801385,
|
| 1776 |
+
"grad_norm": 51.083431243896484,
|
| 1777 |
+
"learning_rate": 4.745256043670393e-05,
|
| 1778 |
+
"loss": 0.9675,
|
| 1779 |
+
"step": 2460
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 1.9014626635873748,
|
| 1783 |
+
"grad_norm": 59.243743896484375,
|
| 1784 |
+
"learning_rate": 4.743956329607487e-05,
|
| 1785 |
+
"loss": 0.7884,
|
| 1786 |
+
"step": 2470
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 1.9091608929946111,
|
| 1790 |
+
"grad_norm": 25.169973373413086,
|
| 1791 |
+
"learning_rate": 4.7426566155445804e-05,
|
| 1792 |
+
"loss": 1.0393,
|
| 1793 |
+
"step": 2480
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.9168591224018474,
|
| 1797 |
+
"grad_norm": 19.93684196472168,
|
| 1798 |
+
"learning_rate": 4.7413569014816746e-05,
|
| 1799 |
+
"loss": 1.0029,
|
| 1800 |
+
"step": 2490
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 1.924557351809084,
|
| 1804 |
+
"grad_norm": 14.145678520202637,
|
| 1805 |
+
"learning_rate": 4.740057187418768e-05,
|
| 1806 |
+
"loss": 1.0533,
|
| 1807 |
+
"step": 2500
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 1.924557351809084,
|
| 1811 |
+
"eval_accuracy": 0.5569668976135489,
|
| 1812 |
+
"eval_f1": 0.5318009748153606,
|
| 1813 |
+
"eval_loss": 0.9681846499443054,
|
| 1814 |
+
"eval_precision": 0.5749243005518604,
|
| 1815 |
+
"eval_recall": 0.5569668976135489,
|
| 1816 |
+
"eval_runtime": 8.0276,
|
| 1817 |
+
"eval_samples_per_second": 323.634,
|
| 1818 |
+
"eval_steps_per_second": 40.485,
|
| 1819 |
+
"step": 2500
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 1.9322555812163202,
|
| 1823 |
+
"grad_norm": 15.758938789367676,
|
| 1824 |
+
"learning_rate": 4.738757473355862e-05,
|
| 1825 |
+
"loss": 0.8101,
|
| 1826 |
+
"step": 2510
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 1.9399538106235565,
|
| 1830 |
+
"grad_norm": 13.67740535736084,
|
| 1831 |
+
"learning_rate": 4.737457759292956e-05,
|
| 1832 |
+
"loss": 0.7461,
|
| 1833 |
+
"step": 2520
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 1.947652040030793,
|
| 1837 |
+
"grad_norm": 15.8422269821167,
|
| 1838 |
+
"learning_rate": 4.73615804523005e-05,
|
| 1839 |
+
"loss": 1.0919,
|
| 1840 |
+
"step": 2530
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 1.9553502694380294,
|
| 1844 |
+
"grad_norm": 150.54856872558594,
|
| 1845 |
+
"learning_rate": 4.7348583311671434e-05,
|
| 1846 |
+
"loss": 0.8516,
|
| 1847 |
+
"step": 2540
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 1.9630484988452657,
|
| 1851 |
+
"grad_norm": 73.50398254394531,
|
| 1852 |
+
"learning_rate": 4.7335586171042376e-05,
|
| 1853 |
+
"loss": 0.9784,
|
| 1854 |
+
"step": 2550
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 1.970746728252502,
|
| 1858 |
+
"grad_norm": 16.76059913635254,
|
| 1859 |
+
"learning_rate": 4.732258903041331e-05,
|
| 1860 |
+
"loss": 0.8167,
|
| 1861 |
+
"step": 2560
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 1.9784449576597383,
|
| 1865 |
+
"grad_norm": 29.237688064575195,
|
| 1866 |
+
"learning_rate": 4.730959188978425e-05,
|
| 1867 |
+
"loss": 1.0536,
|
| 1868 |
+
"step": 2570
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 1.9861431870669746,
|
| 1872 |
+
"grad_norm": 21.804590225219727,
|
| 1873 |
+
"learning_rate": 4.729659474915519e-05,
|
| 1874 |
+
"loss": 0.9558,
|
| 1875 |
+
"step": 2580
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 1.9938414164742109,
|
| 1879 |
+
"grad_norm": 21.874589920043945,
|
| 1880 |
+
"learning_rate": 4.728359760852612e-05,
|
| 1881 |
+
"loss": 0.9748,
|
| 1882 |
+
"step": 2590
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 2.001539645881447,
|
| 1886 |
+
"grad_norm": 32.83283996582031,
|
| 1887 |
+
"learning_rate": 4.7270600467897064e-05,
|
| 1888 |
+
"loss": 1.0102,
|
| 1889 |
+
"step": 2600
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.0092378752886835,
|
| 1893 |
+
"grad_norm": 16.153419494628906,
|
| 1894 |
+
"learning_rate": 4.7257603327268e-05,
|
| 1895 |
+
"loss": 0.8345,
|
| 1896 |
+
"step": 2610
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 2.0169361046959198,
|
| 1900 |
+
"grad_norm": 54.94171142578125,
|
| 1901 |
+
"learning_rate": 4.724460618663894e-05,
|
| 1902 |
+
"loss": 0.8617,
|
| 1903 |
+
"step": 2620
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 2.024634334103156,
|
| 1907 |
+
"grad_norm": 39.80078887939453,
|
| 1908 |
+
"learning_rate": 4.7231609046009876e-05,
|
| 1909 |
+
"loss": 0.8463,
|
| 1910 |
+
"step": 2630
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 2.032332563510393,
|
| 1914 |
+
"grad_norm": 21.685455322265625,
|
| 1915 |
+
"learning_rate": 4.721861190538082e-05,
|
| 1916 |
+
"loss": 0.8706,
|
| 1917 |
+
"step": 2640
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 2.040030792917629,
|
| 1921 |
+
"grad_norm": 26.26354217529297,
|
| 1922 |
+
"learning_rate": 4.720561476475175e-05,
|
| 1923 |
+
"loss": 0.8641,
|
| 1924 |
+
"step": 2650
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 2.0477290223248654,
|
| 1928 |
+
"grad_norm": 139.40414428710938,
|
| 1929 |
+
"learning_rate": 4.7192617624122694e-05,
|
| 1930 |
+
"loss": 0.7537,
|
| 1931 |
+
"step": 2660
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 2.0554272517321017,
|
| 1935 |
+
"grad_norm": 49.84522247314453,
|
| 1936 |
+
"learning_rate": 4.7179620483493636e-05,
|
| 1937 |
+
"loss": 0.7604,
|
| 1938 |
+
"step": 2670
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 2.063125481139338,
|
| 1942 |
+
"grad_norm": 17.73415756225586,
|
| 1943 |
+
"learning_rate": 4.716662334286457e-05,
|
| 1944 |
+
"loss": 0.7963,
|
| 1945 |
+
"step": 2680
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 2.0708237105465743,
|
| 1949 |
+
"grad_norm": 16.671457290649414,
|
| 1950 |
+
"learning_rate": 4.715362620223551e-05,
|
| 1951 |
+
"loss": 0.8546,
|
| 1952 |
+
"step": 2690
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 2.0785219399538106,
|
| 1956 |
+
"grad_norm": 15.332406044006348,
|
| 1957 |
+
"learning_rate": 4.714062906160645e-05,
|
| 1958 |
+
"loss": 0.784,
|
| 1959 |
+
"step": 2700
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 2.086220169361047,
|
| 1963 |
+
"grad_norm": 76.87459564208984,
|
| 1964 |
+
"learning_rate": 4.712763192097739e-05,
|
| 1965 |
+
"loss": 0.9221,
|
| 1966 |
+
"step": 2710
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 2.0939183987682832,
|
| 1970 |
+
"grad_norm": 21.800127029418945,
|
| 1971 |
+
"learning_rate": 4.7114634780348324e-05,
|
| 1972 |
+
"loss": 1.0555,
|
| 1973 |
+
"step": 2720
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 2.1016166281755195,
|
| 1977 |
+
"grad_norm": 31.496492385864258,
|
| 1978 |
+
"learning_rate": 4.7101637639719266e-05,
|
| 1979 |
+
"loss": 0.8673,
|
| 1980 |
+
"step": 2730
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 2.109314857582756,
|
| 1984 |
+
"grad_norm": 9.187371253967285,
|
| 1985 |
+
"learning_rate": 4.70886404990902e-05,
|
| 1986 |
+
"loss": 0.9217,
|
| 1987 |
+
"step": 2740
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 2.117013086989992,
|
| 1991 |
+
"grad_norm": 19.336244583129883,
|
| 1992 |
+
"learning_rate": 4.707564335846114e-05,
|
| 1993 |
+
"loss": 0.8739,
|
| 1994 |
+
"step": 2750
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 2.1247113163972284,
|
| 1998 |
+
"grad_norm": 28.658281326293945,
|
| 1999 |
+
"learning_rate": 4.706264621783208e-05,
|
| 2000 |
+
"loss": 0.8937,
|
| 2001 |
+
"step": 2760
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 2.132409545804465,
|
| 2005 |
+
"grad_norm": 32.35395431518555,
|
| 2006 |
+
"learning_rate": 4.704964907720302e-05,
|
| 2007 |
+
"loss": 0.9527,
|
| 2008 |
+
"step": 2770
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 2.1401077752117015,
|
| 2012 |
+
"grad_norm": 39.60884094238281,
|
| 2013 |
+
"learning_rate": 4.7036651936573955e-05,
|
| 2014 |
+
"loss": 0.9266,
|
| 2015 |
+
"step": 2780
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 2.147806004618938,
|
| 2019 |
+
"grad_norm": 16.985599517822266,
|
| 2020 |
+
"learning_rate": 4.7023654795944896e-05,
|
| 2021 |
+
"loss": 0.9875,
|
| 2022 |
+
"step": 2790
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 2.155504234026174,
|
| 2026 |
+
"grad_norm": 14.943406105041504,
|
| 2027 |
+
"learning_rate": 4.701065765531583e-05,
|
| 2028 |
+
"loss": 0.8371,
|
| 2029 |
+
"step": 2800
|
| 2030 |
+
},
|
| 2031 |
+
{
|
| 2032 |
+
"epoch": 2.1632024634334104,
|
| 2033 |
+
"grad_norm": 18.414188385009766,
|
| 2034 |
+
"learning_rate": 4.699766051468677e-05,
|
| 2035 |
+
"loss": 0.8526,
|
| 2036 |
+
"step": 2810
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 2.1709006928406467,
|
| 2040 |
+
"grad_norm": 12.915871620178223,
|
| 2041 |
+
"learning_rate": 4.698466337405771e-05,
|
| 2042 |
+
"loss": 0.9252,
|
| 2043 |
+
"step": 2820
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 2.178598922247883,
|
| 2047 |
+
"grad_norm": 38.03502655029297,
|
| 2048 |
+
"learning_rate": 4.697166623342864e-05,
|
| 2049 |
+
"loss": 0.8437,
|
| 2050 |
+
"step": 2830
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 2.1862971516551193,
|
| 2054 |
+
"grad_norm": 163.22010803222656,
|
| 2055 |
+
"learning_rate": 4.6958669092799585e-05,
|
| 2056 |
+
"loss": 0.8906,
|
| 2057 |
+
"step": 2840
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 2.1939953810623556,
|
| 2061 |
+
"grad_norm": 15.631921768188477,
|
| 2062 |
+
"learning_rate": 4.694567195217052e-05,
|
| 2063 |
+
"loss": 0.9374,
|
| 2064 |
+
"step": 2850
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 2.201693610469592,
|
| 2068 |
+
"grad_norm": 91.86471557617188,
|
| 2069 |
+
"learning_rate": 4.693267481154146e-05,
|
| 2070 |
+
"loss": 0.895,
|
| 2071 |
+
"step": 2860
|
| 2072 |
+
},
|
| 2073 |
+
{
|
| 2074 |
+
"epoch": 2.209391839876828,
|
| 2075 |
+
"grad_norm": 43.340171813964844,
|
| 2076 |
+
"learning_rate": 4.69196776709124e-05,
|
| 2077 |
+
"loss": 0.9779,
|
| 2078 |
+
"step": 2870
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 2.2170900692840645,
|
| 2082 |
+
"grad_norm": 17.621532440185547,
|
| 2083 |
+
"learning_rate": 4.6906680530283345e-05,
|
| 2084 |
+
"loss": 0.9628,
|
| 2085 |
+
"step": 2880
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 2.2247882986913012,
|
| 2089 |
+
"grad_norm": 22.588891983032227,
|
| 2090 |
+
"learning_rate": 4.689368338965428e-05,
|
| 2091 |
+
"loss": 0.8364,
|
| 2092 |
+
"step": 2890
|
| 2093 |
+
},
|
| 2094 |
+
{
|
| 2095 |
+
"epoch": 2.2324865280985375,
|
| 2096 |
+
"grad_norm": 17.839075088500977,
|
| 2097 |
+
"learning_rate": 4.688068624902522e-05,
|
| 2098 |
+
"loss": 0.8933,
|
| 2099 |
+
"step": 2900
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 2.240184757505774,
|
| 2103 |
+
"grad_norm": 63.97893524169922,
|
| 2104 |
+
"learning_rate": 4.6867689108396157e-05,
|
| 2105 |
+
"loss": 0.8879,
|
| 2106 |
+
"step": 2910
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 2.24788298691301,
|
| 2110 |
+
"grad_norm": 25.88020133972168,
|
| 2111 |
+
"learning_rate": 4.685469196776709e-05,
|
| 2112 |
+
"loss": 1.1249,
|
| 2113 |
+
"step": 2920
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 2.2555812163202464,
|
| 2117 |
+
"grad_norm": 61.59385299682617,
|
| 2118 |
+
"learning_rate": 4.684169482713803e-05,
|
| 2119 |
+
"loss": 0.9538,
|
| 2120 |
+
"step": 2930
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 2.2632794457274827,
|
| 2124 |
+
"grad_norm": 313.2900390625,
|
| 2125 |
+
"learning_rate": 4.682869768650897e-05,
|
| 2126 |
+
"loss": 0.875,
|
| 2127 |
+
"step": 2940
|
| 2128 |
+
},
|
| 2129 |
+
{
|
| 2130 |
+
"epoch": 2.270977675134719,
|
| 2131 |
+
"grad_norm": 24.295677185058594,
|
| 2132 |
+
"learning_rate": 4.681570054587991e-05,
|
| 2133 |
+
"loss": 1.0275,
|
| 2134 |
+
"step": 2950
|
| 2135 |
+
},
|
| 2136 |
+
{
|
| 2137 |
+
"epoch": 2.2786759045419553,
|
| 2138 |
+
"grad_norm": 20.744089126586914,
|
| 2139 |
+
"learning_rate": 4.6802703405250845e-05,
|
| 2140 |
+
"loss": 0.9133,
|
| 2141 |
+
"step": 2960
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 2.2863741339491916,
|
| 2145 |
+
"grad_norm": 52.287078857421875,
|
| 2146 |
+
"learning_rate": 4.678970626462179e-05,
|
| 2147 |
+
"loss": 0.7897,
|
| 2148 |
+
"step": 2970
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 2.294072363356428,
|
| 2152 |
+
"grad_norm": 21.226224899291992,
|
| 2153 |
+
"learning_rate": 4.677670912399272e-05,
|
| 2154 |
+
"loss": 0.9604,
|
| 2155 |
+
"step": 2980
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 2.3017705927636642,
|
| 2159 |
+
"grad_norm": 12.787373542785645,
|
| 2160 |
+
"learning_rate": 4.6763711983363663e-05,
|
| 2161 |
+
"loss": 0.8511,
|
| 2162 |
+
"step": 2990
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 2.3094688221709005,
|
| 2166 |
+
"grad_norm": 22.66034507751465,
|
| 2167 |
+
"learning_rate": 4.67507148427346e-05,
|
| 2168 |
+
"loss": 1.0062,
|
| 2169 |
+
"step": 3000
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 2.3094688221709005,
|
| 2173 |
+
"eval_accuracy": 0.588144726712856,
|
| 2174 |
+
"eval_f1": 0.5671167090208166,
|
| 2175 |
+
"eval_loss": 0.9636054039001465,
|
| 2176 |
+
"eval_precision": 0.5876103570700653,
|
| 2177 |
+
"eval_recall": 0.588144726712856,
|
| 2178 |
+
"eval_runtime": 7.9352,
|
| 2179 |
+
"eval_samples_per_second": 327.401,
|
| 2180 |
+
"eval_steps_per_second": 40.957,
|
| 2181 |
+
"step": 3000
|
| 2182 |
+
}
|
| 2183 |
+
],
|
| 2184 |
+
"logging_steps": 10,
|
| 2185 |
+
"max_steps": 38970,
|
| 2186 |
+
"num_input_tokens_seen": 0,
|
| 2187 |
+
"num_train_epochs": 30,
|
| 2188 |
+
"save_steps": 500,
|
| 2189 |
+
"total_flos": 277929582871200.0,
|
| 2190 |
+
"train_batch_size": 8,
|
| 2191 |
+
"trial_name": null,
|
| 2192 |
+
"trial_params": null
|
| 2193 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06ab9ec733999b8b864e080564fa5a0cc3543db4644af91581f40c86936c1d22
|
| 3 |
+
size 4920
|