Text Generation
English
Generated from Trainer
conversational
Eval Results
zxrzxr commited on
Commit
696de72
1 Parent(s): 9092aeb

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
README.md ADDED
@@ -0,0 +1,414 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - generated_from_trainer
5
+ license: mit
6
+ datasets:
7
+ - HuggingFaceH4/ultrachat_200k
8
+ - HuggingFaceH4/ultrafeedback_binarized
9
+ language:
10
+ - en
11
+ base_model: mistralai/Mistral-7B-v0.1
12
+ widget:
13
+ - text: "<|system|>\nYou are a pirate chatbot who always responds with Arr!</s>\n<|user|>\nThere's a llama on my lawn, how can I get rid of him?</s>\n<|assistant|>\n"
14
+ output:
15
+ text: "Arr! 'Tis a puzzlin' matter, me hearty! A llama on yer lawn be a rare sight, but I've got a plan that might help ye get rid of 'im. Ye'll need to gather some carrots and hay, and then lure the llama away with the promise of a tasty treat. Once he's gone, ye can clean up yer lawn and enjoy the peace and quiet once again. But beware, me hearty, for there may be more llamas where that one came from! Arr!"
16
+ pipeline_tag: text-generation
17
+ model-index:
18
+ - name: zephyr-7b-beta
19
+ results:
20
+ # AI2 Reasoning Challenge (25-Shot)
21
+ - task:
22
+ type: text-generation
23
+ name: Text Generation
24
+ dataset:
25
+ name: AI2 Reasoning Challenge (25-Shot)
26
+ type: ai2_arc
27
+ config: ARC-Challenge
28
+ split: test
29
+ args:
30
+ num_few_shot: 25
31
+ metrics:
32
+ - type: acc_norm
33
+ name: normalized accuracy
34
+ value: 0.6203071672354948
35
+ source:
36
+ name: Open LLM Leaderboard
37
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
38
+
39
+ # HellaSwag (10-shot)
40
+ - task:
41
+ type: text-generation
42
+ name: Text Generation
43
+ dataset:
44
+ name: HellaSwag (10-Shot)
45
+ type: hellaswag
46
+ split: validation
47
+ args:
48
+ num_few_shot: 10
49
+ metrics:
50
+ - type: acc_norm
51
+ name: normalized accuracy
52
+ value: 0.8435570603465445
53
+ source:
54
+ name: Open LLM Leaderboard
55
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
56
+
57
+ # DROP (3-shot)
58
+ - task:
59
+ type: text-generation
60
+ name: Text Generation
61
+ dataset:
62
+ name: Drop (3-Shot)
63
+ type: drop
64
+ split: validation
65
+ args:
66
+ num_few_shot: 3
67
+ metrics:
68
+ - type: f1
69
+ name: f1 score
70
+ value: 0.09662437080536909
71
+ source:
72
+ name: Open LLM Leaderboard
73
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
74
+
75
+ # TruthfulQA (0-shot)
76
+ - task:
77
+ type: text-generation
78
+ name: Text Generation
79
+ dataset:
80
+ name: TruthfulQA (0-shot)
81
+ type: truthful_qa
82
+ config: multiple_choice
83
+ split: validation
84
+ args:
85
+ num_few_shot: 0
86
+ metrics:
87
+ - type: mc2
88
+ value: 0.5744916942762855
89
+ source:
90
+ name: Open LLM Leaderboard
91
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
92
+
93
+ # GSM8k (5-shot)
94
+ - task:
95
+ type: text-generation
96
+ name: Text Generation
97
+ dataset:
98
+ name: GSM8k (5-shot)
99
+ type: gsm8k
100
+ config: main
101
+ split: test
102
+ args:
103
+ num_few_shot: 5
104
+ metrics:
105
+ - type: acc
106
+ name: accuracy
107
+ value: 0.12736921910538287
108
+ source:
109
+ name: Open LLM Leaderboard
110
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
111
+
112
+ # MMLU (5-Shot)
113
+ - task:
114
+ type: text-generation
115
+ name: Text Generation
116
+ dataset:
117
+ name: MMLU (5-Shot)
118
+ type: cais/mmlu
119
+ config: all
120
+ split: test
121
+ args:
122
+ num_few_shot: 5
123
+ metrics:
124
+ - type: acc
125
+ name: accuracy
126
+ value: 0.6107
127
+ source:
128
+ name: Open LLM Leaderboard
129
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
130
+
131
+ # Winogrande (5-shot)
132
+ - task:
133
+ type: text-generation
134
+ name: Text Generation
135
+ dataset:
136
+ name: Winogrande (5-shot)
137
+ type: winogrande
138
+ config: winogrande_xl
139
+ split: validation
140
+ args:
141
+ num_few_shot: 5
142
+ metrics:
143
+ - type: acc
144
+ name: accuracy
145
+ value: 0.7774269928966061
146
+ source:
147
+ name: Open LLM Leaderboard
148
+ url: https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta_public
149
+
150
+ # AlpacaEval (taken from model card)
151
+ - task:
152
+ type: text-generation
153
+ name: Text Generation
154
+ dataset:
155
+ name: AlpacaEval
156
+ type: tatsu-lab/alpaca_eval
157
+ metrics:
158
+ - type: unknown
159
+ name: win rate
160
+ value: 0.9060
161
+ source:
162
+ url: https://tatsu-lab.github.io/alpaca_eval/
163
+
164
+ # MT-Bench (taken from model card)
165
+ - task:
166
+ type: text-generation
167
+ name: Text Generation
168
+ dataset:
169
+ name: MT-Bench
170
+ type: unknown
171
+ metrics:
172
+ - type: unknown
173
+ name: score
174
+ value: 7.34
175
+ source:
176
+ url: https://huggingface.co/spaces/lmsys/mt-bench
177
+ ---
178
+
179
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
180
+ should probably proofread and complete it, then remove this comment. -->
181
+
182
+ <img src="https://huggingface.co/HuggingFaceH4/zephyr-7b-alpha/resolve/main/thumbnail.png" alt="Zephyr Logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/>
183
+
184
+
185
+ # Model Card for Zephyr 7B β
186
+
187
+ This is a wasm quantized version of zephyr 7b beta.
188
+ Compiled by using [mlc-llm](https://github.com/mlc-ai/mlc-llm)
189
+
190
+ Zephyr is a series of language models that are trained to act as helpful assistants. Zephyr-7B-β is the second model in the series, and is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) that was trained on on a mix of publicly available, synthetic datasets using [Direct Preference Optimization (DPO)](https://arxiv.org/abs/2305.18290). We found that removing the in-built alignment of these datasets boosted performance on [MT Bench](https://huggingface.co/spaces/lmsys/mt-bench) and made the model more helpful. However, this means that model is likely to generate problematic text when prompted to do so. You can find more details in the [technical report](https://arxiv.org/abs/2310.16944).
191
+
192
+
193
+ ## Model description
194
+
195
+ - **Model type:** A 7B parameter GPT-like model fine-tuned on a mix of publicly available, synthetic datasets.
196
+ - **Language(s) (NLP):** Primarily English
197
+ - **License:** MIT
198
+ - **Finetuned from model:** [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)
199
+
200
+ ### Model Sources
201
+
202
+ <!-- Provide the basic links for the model. -->
203
+
204
+ - **Repository:** https://github.com/huggingface/alignment-handbook
205
+ - **Demo:** https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat
206
+ - **Chatbot Arena:** Evaluate Zephyr 7B against 10+ LLMs in the LMSYS arena: http://arena.lmsys.org
207
+
208
+ ## Performance
209
+
210
+ At the time of release, Zephyr-7B-β is the highest ranked 7B chat model on the [MT-Bench](https://huggingface.co/spaces/lmsys/mt-bench) and [AlpacaEval](https://tatsu-lab.github.io/alpaca_eval/) benchmarks:
211
+
212
+ | Model | Size | Alignment | MT-Bench (score) | AlpacaEval (win rate %) |
213
+ |-------------|-----|----|---------------|--------------|
214
+ | StableLM-Tuned-α | 7B| dSFT |2.75| -|
215
+ | MPT-Chat | 7B |dSFT |5.42| -|
216
+ | Xwin-LMv0.1 | 7B| dPPO| 6.19| 87.83|
217
+ | Mistral-Instructv0.1 | 7B| - | 6.84 |-|
218
+ | Zephyr-7b-α |7B| dDPO| 6.88| -|
219
+ | **Zephyr-7b-β** 🪁 | **7B** | **dDPO** | **7.34** | **90.60** |
220
+ | Falcon-Instruct | 40B |dSFT |5.17 |45.71|
221
+ | Guanaco | 65B | SFT |6.41| 71.80|
222
+ | Llama2-Chat | 70B |RLHF |6.86| 92.66|
223
+ | Vicuna v1.3 | 33B |dSFT |7.12 |88.99|
224
+ | WizardLM v1.0 | 70B |dSFT |7.71 |-|
225
+ | Xwin-LM v0.1 | 70B |dPPO |- |95.57|
226
+ | GPT-3.5-turbo | - |RLHF |7.94 |89.37|
227
+ | Claude 2 | - |RLHF |8.06| 91.36|
228
+ | GPT-4 | -| RLHF |8.99| 95.28|
229
+
230
+ In particular, on several categories of MT-Bench, Zephyr-7B-β has strong performance compared to larger open models like Llama2-Chat-70B:
231
+
232
+ ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6200d0a443eb0913fa2df7cc/raxvt5ma16d7T23my34WC.png)
233
+
234
+ However, on more complex tasks like coding and mathematics, Zephyr-7B-β lags behind proprietary models and more research is needed to close the gap.
235
+
236
+
237
+ ## Intended uses & limitations
238
+
239
+ The model was initially fine-tuned on a filtered and preprocessed of the [`UltraChat`](https://huggingface.co/datasets/stingning/ultrachat) dataset, which contains a diverse range of synthetic dialogues generated by ChatGPT.
240
+ We then further aligned the model with [🤗 TRL's](https://github.com/huggingface/trl) `DPOTrainer` on the [openbmb/UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) dataset, which contains 64k prompts and model completions that are ranked by GPT-4. As a result, the model can be used for chat and you can check out our [demo](https://huggingface.co/spaces/HuggingFaceH4/zephyr-chat) to test its capabilities.
241
+
242
+ You can find the datasets used for training Zephyr-7B-β [here](https://huggingface.co/collections/HuggingFaceH4/zephyr-7b-6538c6d6d5ddd1cbb1744a66)
243
+
244
+ Here's how you can run the model using the `pipeline()` function from 🤗 Transformers:
245
+
246
+ ```python
247
+ # Install transformers from source - only needed for versions <= v4.34
248
+ # pip install git+https://github.com/huggingface/transformers.git
249
+ # pip install accelerate
250
+
251
+ import torch
252
+ from transformers import pipeline
253
+
254
+ pipe = pipeline("text-generation", model="HuggingFaceH4/zephyr-7b-beta", torch_dtype=torch.bfloat16, device_map="auto")
255
+
256
+ # We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
257
+ messages = [
258
+ {
259
+ "role": "system",
260
+ "content": "You are a friendly chatbot who always responds in the style of a pirate",
261
+ },
262
+ {"role": "user", "content": "How many helicopters can a human eat in one sitting?"},
263
+ ]
264
+ prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
265
+ outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
266
+ print(outputs[0]["generated_text"])
267
+ # <|system|>
268
+ # You are a friendly chatbot who always responds in the style of a pirate.</s>
269
+ # <|user|>
270
+ # How many helicopters can a human eat in one sitting?</s>
271
+ # <|assistant|>
272
+ # Ah, me hearty matey! But yer question be a puzzler! A human cannot eat a helicopter in one sitting, as helicopters are not edible. They be made of metal, plastic, and other materials, not food!
273
+ ```
274
+
275
+ ## Bias, Risks, and Limitations
276
+
277
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
278
+
279
+ Zephyr-7B-β has not been aligned to human preferences for safety within the RLHF phase or deployed with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so).
280
+ It is also unknown what the size and composition of the corpus was used to train the base model (`mistralai/Mistral-7B-v0.1`), however it is likely to have included a mix of Web data and technical sources like books and code. See the [Falcon 180B model card](https://huggingface.co/tiiuae/falcon-180B#training-data) for an example of this.
281
+
282
+
283
+ ## Training and evaluation data
284
+
285
+ During DPO training, this model achieves the following results on the evaluation set:
286
+
287
+ - Loss: 0.7496
288
+ - Rewards/chosen: -4.5221
289
+ - Rewards/rejected: -8.3184
290
+ - Rewards/accuracies: 0.7812
291
+ - Rewards/margins: 3.7963
292
+ - Logps/rejected: -340.1541
293
+ - Logps/chosen: -299.4561
294
+ - Logits/rejected: -2.3081
295
+ - Logits/chosen: -2.3531
296
+
297
+
298
+ ### Training hyperparameters
299
+
300
+ The following hyperparameters were used during training:
301
+ - learning_rate: 5e-07
302
+ - train_batch_size: 2
303
+ - eval_batch_size: 4
304
+ - seed: 42
305
+ - distributed_type: multi-GPU
306
+ - num_devices: 16
307
+ - total_train_batch_size: 32
308
+ - total_eval_batch_size: 64
309
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
310
+ - lr_scheduler_type: linear
311
+ - lr_scheduler_warmup_ratio: 0.1
312
+ - num_epochs: 3.0
313
+
314
+ ### Training results
315
+
316
+ The table below shows the full set of DPO training metrics:
317
+
318
+
319
+ | Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
320
+ |:-------------:|:-----:|:----:|:---------------:|:--------------:|:----------------:|:------------------:|:---------------:|:--------------:|:------------:|:---------------:|:-------------:|
321
+ | 0.6284 | 0.05 | 100 | 0.6098 | 0.0425 | -0.1872 | 0.7344 | 0.2297 | -258.8416 | -253.8099 | -2.7976 | -2.8234 |
322
+ | 0.4908 | 0.1 | 200 | 0.5426 | -0.0279 | -0.6842 | 0.75 | 0.6563 | -263.8124 | -254.5145 | -2.7719 | -2.7960 |
323
+ | 0.5264 | 0.15 | 300 | 0.5324 | 0.0414 | -0.9793 | 0.7656 | 1.0207 | -266.7627 | -253.8209 | -2.7892 | -2.8122 |
324
+ | 0.5536 | 0.21 | 400 | 0.4957 | -0.0185 | -1.5276 | 0.7969 | 1.5091 | -272.2460 | -254.4203 | -2.8542 | -2.8764 |
325
+ | 0.5362 | 0.26 | 500 | 0.5031 | -0.2630 | -1.5917 | 0.7812 | 1.3287 | -272.8869 | -256.8653 | -2.8702 | -2.8958 |
326
+ | 0.5966 | 0.31 | 600 | 0.5963 | -0.2993 | -1.6491 | 0.7812 | 1.3499 | -273.4614 | -257.2279 | -2.8778 | -2.8986 |
327
+ | 0.5014 | 0.36 | 700 | 0.5382 | -0.2859 | -1.4750 | 0.75 | 1.1891 | -271.7204 | -257.0942 | -2.7659 | -2.7869 |
328
+ | 0.5334 | 0.41 | 800 | 0.5677 | -0.4289 | -1.8968 | 0.7969 | 1.4679 | -275.9378 | -258.5242 | -2.7053 | -2.7265 |
329
+ | 0.5251 | 0.46 | 900 | 0.5772 | -0.2116 | -1.3107 | 0.7344 | 1.0991 | -270.0768 | -256.3507 | -2.8463 | -2.8662 |
330
+ | 0.5205 | 0.52 | 1000 | 0.5262 | -0.3792 | -1.8585 | 0.7188 | 1.4793 | -275.5552 | -258.0276 | -2.7893 | -2.7979 |
331
+ | 0.5094 | 0.57 | 1100 | 0.5433 | -0.6279 | -1.9368 | 0.7969 | 1.3089 | -276.3377 | -260.5136 | -2.7453 | -2.7536 |
332
+ | 0.5837 | 0.62 | 1200 | 0.5349 | -0.3780 | -1.9584 | 0.7656 | 1.5804 | -276.5542 | -258.0154 | -2.7643 | -2.7756 |
333
+ | 0.5214 | 0.67 | 1300 | 0.5732 | -1.0055 | -2.2306 | 0.7656 | 1.2251 | -279.2761 | -264.2903 | -2.6986 | -2.7113 |
334
+ | 0.6914 | 0.72 | 1400 | 0.5137 | -0.6912 | -2.1775 | 0.7969 | 1.4863 | -278.7448 | -261.1467 | -2.7166 | -2.7275 |
335
+ | 0.4655 | 0.77 | 1500 | 0.5090 | -0.7987 | -2.2930 | 0.7031 | 1.4943 | -279.8999 | -262.2220 | -2.6651 | -2.6838 |
336
+ | 0.5731 | 0.83 | 1600 | 0.5312 | -0.8253 | -2.3520 | 0.7812 | 1.5268 | -280.4902 | -262.4876 | -2.6543 | -2.6728 |
337
+ | 0.5233 | 0.88 | 1700 | 0.5206 | -0.4573 | -2.0951 | 0.7812 | 1.6377 | -277.9205 | -258.8084 | -2.6870 | -2.7097 |
338
+ | 0.5593 | 0.93 | 1800 | 0.5231 | -0.5508 | -2.2000 | 0.7969 | 1.6492 | -278.9703 | -259.7433 | -2.6221 | -2.6519 |
339
+ | 0.4967 | 0.98 | 1900 | 0.5290 | -0.5340 | -1.9570 | 0.8281 | 1.4230 | -276.5395 | -259.5749 | -2.6564 | -2.6878 |
340
+ | 0.0921 | 1.03 | 2000 | 0.5368 | -1.1376 | -3.1615 | 0.7812 | 2.0239 | -288.5854 | -265.6111 | -2.6040 | -2.6345 |
341
+ | 0.0733 | 1.08 | 2100 | 0.5453 | -1.1045 | -3.4451 | 0.7656 | 2.3406 | -291.4208 | -265.2799 | -2.6289 | -2.6595 |
342
+ | 0.0972 | 1.14 | 2200 | 0.5571 | -1.6915 | -3.9823 | 0.8125 | 2.2908 | -296.7934 | -271.1505 | -2.6471 | -2.6709 |
343
+ | 0.1058 | 1.19 | 2300 | 0.5789 | -1.0621 | -3.8941 | 0.7969 | 2.8319 | -295.9106 | -264.8563 | -2.5527 | -2.5798 |
344
+ | 0.2423 | 1.24 | 2400 | 0.5455 | -1.1963 | -3.5590 | 0.7812 | 2.3627 | -292.5599 | -266.1981 | -2.5414 | -2.5784 |
345
+ | 0.1177 | 1.29 | 2500 | 0.5889 | -1.8141 | -4.3942 | 0.7969 | 2.5801 | -300.9120 | -272.3761 | -2.4802 | -2.5189 |
346
+ | 0.1213 | 1.34 | 2600 | 0.5683 | -1.4608 | -3.8420 | 0.8125 | 2.3812 | -295.3901 | -268.8436 | -2.4774 | -2.5207 |
347
+ | 0.0889 | 1.39 | 2700 | 0.5890 | -1.6007 | -3.7337 | 0.7812 | 2.1330 | -294.3068 | -270.2423 | -2.4123 | -2.4522 |
348
+ | 0.0995 | 1.45 | 2800 | 0.6073 | -1.5519 | -3.8362 | 0.8281 | 2.2843 | -295.3315 | -269.7538 | -2.4685 | -2.5050 |
349
+ | 0.1145 | 1.5 | 2900 | 0.5790 | -1.7939 | -4.2876 | 0.8438 | 2.4937 | -299.8461 | -272.1744 | -2.4272 | -2.4674 |
350
+ | 0.0644 | 1.55 | 3000 | 0.5735 | -1.7285 | -4.2051 | 0.8125 | 2.4766 | -299.0209 | -271.5201 | -2.4193 | -2.4574 |
351
+ | 0.0798 | 1.6 | 3100 | 0.5537 | -1.7226 | -4.2850 | 0.8438 | 2.5624 | -299.8200 | -271.4610 | -2.5367 | -2.5696 |
352
+ | 0.1013 | 1.65 | 3200 | 0.5575 | -1.5715 | -3.9813 | 0.875 | 2.4098 | -296.7825 | -269.9498 | -2.4926 | -2.5267 |
353
+ | 0.1254 | 1.7 | 3300 | 0.5905 | -1.6412 | -4.4703 | 0.8594 | 2.8291 | -301.6730 | -270.6473 | -2.5017 | -2.5340 |
354
+ | 0.085 | 1.76 | 3400 | 0.6133 | -1.9159 | -4.6760 | 0.8438 | 2.7601 | -303.7296 | -273.3941 | -2.4614 | -2.4960 |
355
+ | 0.065 | 1.81 | 3500 | 0.6074 | -1.8237 | -4.3525 | 0.8594 | 2.5288 | -300.4951 | -272.4724 | -2.4597 | -2.5004 |
356
+ | 0.0755 | 1.86 | 3600 | 0.5836 | -1.9252 | -4.4005 | 0.8125 | 2.4753 | -300.9748 | -273.4872 | -2.4327 | -2.4716 |
357
+ | 0.0746 | 1.91 | 3700 | 0.5789 | -1.9280 | -4.4906 | 0.8125 | 2.5626 | -301.8762 | -273.5149 | -2.4686 | -2.5115 |
358
+ | 0.1348 | 1.96 | 3800 | 0.6015 | -1.8658 | -4.2428 | 0.8281 | 2.3769 | -299.3976 | -272.8936 | -2.4943 | -2.5393 |
359
+ | 0.0217 | 2.01 | 3900 | 0.6122 | -2.3335 | -4.9229 | 0.8281 | 2.5894 | -306.1988 | -277.5699 | -2.4841 | -2.5272 |
360
+ | 0.0219 | 2.07 | 4000 | 0.6522 | -2.9890 | -6.0164 | 0.8281 | 3.0274 | -317.1334 | -284.1248 | -2.4105 | -2.4545 |
361
+ | 0.0119 | 2.12 | 4100 | 0.6922 | -3.4777 | -6.6749 | 0.7969 | 3.1972 | -323.7187 | -289.0121 | -2.4272 | -2.4699 |
362
+ | 0.0153 | 2.17 | 4200 | 0.6993 | -3.2406 | -6.6775 | 0.7969 | 3.4369 | -323.7453 | -286.6413 | -2.4047 | -2.4465 |
363
+ | 0.011 | 2.22 | 4300 | 0.7178 | -3.7991 | -7.4397 | 0.7656 | 3.6406 | -331.3667 | -292.2260 | -2.3843 | -2.4290 |
364
+ | 0.0072 | 2.27 | 4400 | 0.6840 | -3.3269 | -6.8021 | 0.8125 | 3.4752 | -324.9908 | -287.5042 | -2.4095 | -2.4536 |
365
+ | 0.0197 | 2.32 | 4500 | 0.7013 | -3.6890 | -7.3014 | 0.8125 | 3.6124 | -329.9841 | -291.1250 | -2.4118 | -2.4543 |
366
+ | 0.0182 | 2.37 | 4600 | 0.7476 | -3.8994 | -7.5366 | 0.8281 | 3.6372 | -332.3356 | -293.2291 | -2.4163 | -2.4565 |
367
+ | 0.0125 | 2.43 | 4700 | 0.7199 | -4.0560 | -7.5765 | 0.8438 | 3.5204 | -332.7345 | -294.7952 | -2.3699 | -2.4100 |
368
+ | 0.0082 | 2.48 | 4800 | 0.7048 | -3.6613 | -7.1356 | 0.875 | 3.4743 | -328.3255 | -290.8477 | -2.3925 | -2.4303 |
369
+ | 0.0118 | 2.53 | 4900 | 0.6976 | -3.7908 | -7.3152 | 0.8125 | 3.5244 | -330.1224 | -292.1431 | -2.3633 | -2.4047 |
370
+ | 0.0118 | 2.58 | 5000 | 0.7198 | -3.9049 | -7.5557 | 0.8281 | 3.6508 | -332.5271 | -293.2844 | -2.3764 | -2.4194 |
371
+ | 0.006 | 2.63 | 5100 | 0.7506 | -4.2118 | -7.9149 | 0.8125 | 3.7032 | -336.1194 | -296.3530 | -2.3407 | -2.3860 |
372
+ | 0.0143 | 2.68 | 5200 | 0.7408 | -4.2433 | -7.9802 | 0.8125 | 3.7369 | -336.7721 | -296.6682 | -2.3509 | -2.3946 |
373
+ | 0.0057 | 2.74 | 5300 | 0.7552 | -4.3392 | -8.0831 | 0.7969 | 3.7439 | -337.8013 | -297.6275 | -2.3388 | -2.3842 |
374
+ | 0.0138 | 2.79 | 5400 | 0.7404 | -4.2395 | -7.9762 | 0.8125 | 3.7367 | -336.7322 | -296.6304 | -2.3286 | -2.3737 |
375
+ | 0.0079 | 2.84 | 5500 | 0.7525 | -4.4466 | -8.2196 | 0.7812 | 3.7731 | -339.1662 | -298.7007 | -2.3200 | -2.3641 |
376
+ | 0.0077 | 2.89 | 5600 | 0.7520 | -4.5586 | -8.3485 | 0.7969 | 3.7899 | -340.4545 | -299.8206 | -2.3078 | -2.3517 |
377
+ | 0.0094 | 2.94 | 5700 | 0.7527 | -4.5542 | -8.3509 | 0.7812 | 3.7967 | -340.4790 | -299.7773 | -2.3062 | -2.3510 |
378
+ | 0.0054 | 2.99 | 5800 | 0.7520 | -4.5169 | -8.3079 | 0.7812 | 3.7911 | -340.0493 | -299.4038 | -2.3081 | -2.3530 |
379
+
380
+
381
+ ### Framework versions
382
+
383
+ - Transformers 4.35.0.dev0
384
+ - Pytorch 2.0.1+cu118
385
+ - Datasets 2.12.0
386
+ - Tokenizers 0.14.0
387
+
388
+ ## Citation
389
+
390
+ If you find Zephyr-7B-β is useful in your work, please cite it with:
391
+
392
+ ```
393
+ @misc{tunstall2023zephyr,
394
+ title={Zephyr: Direct Distillation of LM Alignment},
395
+ author={Lewis Tunstall and Edward Beeching and Nathan Lambert and Nazneen Rajani and Kashif Rasul and Younes Belkada and Shengyi Huang and Leandro von Werra and Clémentine Fourrier and Nathan Habib and Nathan Sarrazin and Omar Sanseviero and Alexander M. Rush and Thomas Wolf},
396
+ year={2023},
397
+ eprint={2310.16944},
398
+ archivePrefix={arXiv},
399
+ primaryClass={cs.LG}
400
+ }
401
+ ```
402
+ # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
403
+ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_HuggingFaceH4__zephyr-7b-beta)
404
+
405
+ | Metric | Value |
406
+ |-----------------------|---------------------------|
407
+ | Avg. | 52.15 |
408
+ | ARC (25-shot) | 62.03 |
409
+ | HellaSwag (10-shot) | 84.36 |
410
+ | MMLU (5-shot) | 61.07 |
411
+ | TruthfulQA (0-shot) | 57.45 |
412
+ | Winogrande (5-shot) | 77.74 |
413
+ | GSM8K (5-shot) | 12.74 |
414
+ | DROP (3-shot) | 9.66 |
added_tokens.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "</s>": 2,
3
+ "<s>": 1,
4
+ "<unk>": 0
5
+ }
mlc-chat-config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_lib": "zephyr-7b-beta-chatRDM-q4f32_1",
3
+ "local_id": "zephyr-7b-beta-q4f32_1",
4
+ "conv_template": "mistral_default",
5
+ "temperature": 0.7,
6
+ "repetition_penalty": 1.0,
7
+ "top_p": 0.95,
8
+ "mean_gen_len": 128,
9
+ "max_gen_len": 16384,
10
+ "num_shards": 1,
11
+ "use_presharded_weights": false,
12
+ "shift_fill_factor": 0.3,
13
+ "tokenizer_files": [
14
+ "tokenizer.json",
15
+ "added_tokens.json",
16
+ "tokenizer.model"
17
+ ],
18
+ "model_category": "mistral",
19
+ "model_name": "zephyr-7b-beta",
20
+ "vocab_size": 32000,
21
+ "prefill_chunk_size": 4096,
22
+ "sliding_window": 4096
23
+ }
mod_cache_before_build.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d472420ec1330370bdb12a10bed0cbcbc71b39365836e7598b76cef21418b082
3
+ size 32041362
ndarray-cache.json ADDED
The diff for this file is too large to render. See raw diff
 
params_shard_0.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37c67aba48951be930fcafcd0fb3fe3656bbd9e9a3dd0ecfd8962bf53a58ae7d
3
+ size 65536000
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c31437f948f7020611065ab398fccff1108cbcfcdabc37bcdf920a660ed5500
3
+ size 32112640
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a21c06f44bfafa74ac26103e93afb1498effd97db89b2fb6d6bbb8131645d4b7
3
+ size 26214400
params_shard_100.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c36ce1115de518e23a13b0f9f7c09b34f1e45d452426e3647922d59fbadf5143
3
+ size 22052864
params_shard_101.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:631a1e1ce577b3cb72f451b18bf018f6ea63a3437974bf1f15e2083e50969866
3
+ size 58720256
params_shard_102.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e567a24bd7e68c283bf306724e87045c96658bc638baf3e8dc9a9f9112250f4
3
+ size 26214400
params_shard_103.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa512cc3c9f8cb8fd8ecf9c58ec9057a956a2db83ae92de64e1f27fd25536fdc
3
+ size 29360128
params_shard_104.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:389900dafd761a1076ecc51a837ea9d39c8248d9e0f84a2805f5539fff3fe690
3
+ size 22052864
params_shard_105.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a928c53c4126d78e9d3d462d655d40060fcd227d61ea72510e66a74a0275b749
3
+ size 58720256
params_shard_106.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96211089d0f95e3ae21ac52a63b4e7aef3c39274586961641894e4c9ceebe16f
3
+ size 26214400
params_shard_107.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a45f3300f64d14265389785ed9f8b2f17333da4a770ddae69265ac1e6c67d466
3
+ size 29360128
params_shard_108.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91b5727046107dc9d94e56845cd0d3ec42ab2f1931967af0cf82ddaf87f101c8
3
+ size 22052864
params_shard_109.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80e309ade3afd20ac29d3df51a0a7fbca7494d7d6c12b2f4ac8758b8c7d39607
3
+ size 58720256
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b08a683623ecc850564a0745c441506adb52576404fec52e43ef9c2cc46f922b
3
+ size 29360128
params_shard_110.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:caf1ca7fdad52ae4e10d8305dce7071582a1b9192d1872b19ba0aabedf93dbdd
3
+ size 26214400
params_shard_111.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64e230675c03fe3776708944e5e01926ca4446a37b483455cce5700f23b461c5
3
+ size 29360128
params_shard_112.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1b07a17697fd00280783e9a2e7a6babc47754a20e72dbd37a166355cace48626
3
+ size 22052864
params_shard_113.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68f2dfb287bf2628e24f86794b6287262dad651e211bdb63d09db010bcb65b4a
3
+ size 58720256
params_shard_114.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:805f1952ceacc776758b6025e4ca6a414ea325a220399d0024839b6556e79576
3
+ size 26214400
params_shard_115.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:060dfc9380835d89aeada0a9fc04de1d42c4f611d82f7db5222d1dda7c9c333b
3
+ size 29360128
params_shard_116.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:515cabce85f08ed1df13a1d36ff79a12e0749037b35643556cf205ceb921ef03
3
+ size 22052864
params_shard_117.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac694299b6ae19b4f65aaef44cbb6b170acb539aa511663aff081ef5a85aea83
3
+ size 58720256
params_shard_118.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b0faad8c321835deb1a3a1fb286bf0d8f2b219d305e8a5e48bf718bacc1568c
3
+ size 26214400
params_shard_119.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d46224cfca64dc0de596ce47e363e7d2ac70b47120d92579b55473fa11acf94
3
+ size 29360128
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:161918f0364169c97e1f3efce60988fabc7a69170dac25f38be583dc8a54a779
3
+ size 22052864
params_shard_120.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12712b71b9229d2a7cd520d63c1a4b6875304669c648cefe829aab8c5b19a663
3
+ size 22052864
params_shard_121.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2f199431ed2419eef39c52c00b2bea1f70101199ff22448485e8d0496c0c2be2
3
+ size 58720256
params_shard_122.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8d32a150ce3449c4357bebc76f8966e414feca8d39e09879bbbf1f548a53990
3
+ size 26214400
params_shard_123.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c6856de69668883cd473282f4df18bbaa0b8209b22e505f01f9995b0de6238b
3
+ size 29360128
params_shard_124.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55262fec8a7580e58e683d707dd3a77ae71eaf7e5c91a11db6965861d1845b18
3
+ size 22052864
params_shard_125.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c6b0b7b635b319cd1ad7ab2027323a1cf4476a3bd5b7cc4e379e9d3a8ff1453
3
+ size 58720256
params_shard_126.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc068decb257b57a7dee18dcfee6e13c792720b13757b4a3108419f0c6e936c5
3
+ size 26214400
params_shard_127.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e25b38aea0caf52dbcf416eff72d2e2e97b6b0b87718e419fb5aa7b8f4739f55
3
+ size 29360128
params_shard_128.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:344677f90a779cda44797cfbbf2541ab682d43e8e520c9156c1fa8e07b3ab84f
3
+ size 65536000
params_shard_129.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a31315a1ecf8ae2157b77912218aca44749b5513c45022c01d70ca6f350a5b0
3
+ size 22069248
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:745afbc415c6633cf34f152e4f88a19f0adc6af1fdf0acc56836afac39b397db
3
+ size 58720256
params_shard_130.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3d7c5d00723beeb5ca5b38404de413cc8d3711ec8439be77130a8afad0689f03
3
+ size 18481152
params_shard_14.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a5574e7d49778d028bf3bebf3553e95ff7c1df4058ad99d5093e1b5a623b40b
3
+ size 26214400
params_shard_15.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a24b4e96252ae86dc66d8f0cae44dae0fe8d430c96e4d92d5f0de9bc2b0905d
3
+ size 29360128
params_shard_16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51287a06d5d0c20c6d99b0303864ba19de24d4ef54f5c9db82b767ec967a38f4
3
+ size 22052864
params_shard_17.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89df4df2dbe4a6cf2fbeb70a712e65fcd58fa29afae07d6cd5e3bec7ce6807f8
3
+ size 58720256
params_shard_18.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb5c1ba2d1174ea16bf1c38d9e1351b70e7376bdff177a55f8b4a37f2b01cc75
3
+ size 26214400
params_shard_19.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cea0941e14a04d4ec941fd768d17ca2742dea87c7c93c1b1e2a976d07f4b6ed9
3
+ size 29360128
params_shard_2.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:623329f9bebb12d686883795fab9cd211d8df646d687a946070a78784f9e2e27
3
+ size 58720256
params_shard_20.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:082be09805aafc602dc56652d8bcfc8741afa3833a1c28e834e80d357534d5c4
3
+ size 22052864