Hugging Face's logo

Walmart-the-bag
/
Yi-6B-Infinity-Chat

Text Generation
Transformers
PyTorch
Safetensors
llama
conversational
Eval Results
text-generation-inference
Model card Files Community
2
Walmart-the-bag commited on
Commit
4c887d0
•
1 Parent(s): 2b8809e

Upload folder using huggingface_hub

Browse files
Files changed (17) hide show
  1. .ipynb_checkpoints/README-checkpoint.md +55 -0
  2. README.md +618 -0
  3. added_tokens.json +4 -0
  4. all_results.json +7 -0
  5. config.json +27 -0
  6. generation_config.json +10 -0
  7. pytorch_model-00001-of-00002.bin +3 -0
  8. pytorch_model-00002-of-00002.bin +3 -0
  9. pytorch_model.bin.index.json +298 -0
  10. special_tokens_map.json +35 -0
  11. tokenizer.json +0 -0
  12. tokenizer.model +3 -0
  13. tokenizer_config.json +89 -0
  14. train_results.json +7 -0
  15. trainer_log.jsonl +76 -0
  16. trainer_state.json +478 -0
  17. training_args.bin +3 -0
.ipynb_checkpoints/README-checkpoint.md ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ base_model: 01-ai/Yi-6B-Chat
4
+ tags:
5
+ - llama-factory
6
+ - generated_from_trainer
7
+ model-index:
8
+ - name: train_2023-12-23-19-37-58
9
+ results: []
10
+ ---
11
+
12
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
13
+ should probably proofread and complete it, then remove this comment. -->
14
+
15
+ # train_2023-12-23-19-37-58
16
+
17
+ This model is a fine-tuned version of [01-ai/Yi-6B-Chat](https://huggingface.co/01-ai/Yi-6B-Chat) on the alpaca_gpt4_en dataset.
18
+
19
+ ## Model description
20
+
21
+ More information needed
22
+
23
+ ## Intended uses & limitations
24
+
25
+ More information needed
26
+
27
+ ## Training and evaluation data
28
+
29
+ More information needed
30
+
31
+ ## Training procedure
32
+
33
+ ### Training hyperparameters
34
+
35
+ The following hyperparameters were used during training:
36
+ - learning_rate: 5e-05
37
+ - train_batch_size: 4
38
+ - eval_batch_size: 8
39
+ - seed: 42
40
+ - gradient_accumulation_steps: 4
41
+ - total_train_batch_size: 16
42
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
+ - lr_scheduler_type: cosine
44
+ - num_epochs: 3.0
45
+
46
+ ### Training results
47
+
48
+
49
+
50
+ ### Framework versions
51
+
52
+ - Transformers 4.34.1
53
+ - Pytorch 2.1.2+cu121
54
+ - Datasets 2.14.7
55
+ - Tokenizers 0.14.1
README.md CHANGED
@@ -1,3 +1,621 @@
1
  ---
2
  license: cc
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: cc
3
+ base_model: 01-ai/Yi-6B-Chat
4
  ---
5
+
6
+ ## Model description
7
+
8
+ This model is a fine-tuned version of [01-ai/Yi-6B-Chat](https://huggingface.co/01-ai/Yi-6B-Chat) on the alpaca_gpt4_en dataset.
9
+
10
+ # ORIGINAL MODEL CARD:
11
+ ______
12
+
13
+ <div align="center">
14
+
15
+ <picture>
16
+ <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/01-ai/Yi/main/assets/img/Yi_logo_icon_dark.svg" width="200px">
17
+ <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/01-ai/Yi/main/assets/img/Yi_logo_icon_light.svg" width="200px">
18
+ <img alt="specify theme context for images" src="https://raw.githubusercontent.com/01-ai/Yi/main/assets/img/Yi_logo_icon_light.svg">
19
+ </picture>
20
+
21
+ </br>
22
+ </br>
23
+
24
+ <div style="display: inline-block;">
25
+ <a href="https://github.com/01-ai/Yi/actions/workflows/build_docker_image.yml">
26
+ <img src="https://github.com/01-ai/Yi/actions/workflows/build_docker_image.yml/badge.svg">
27
+ </a>
28
+ </div>
29
+
30
+ <div style="display: inline-block;">
31
+ <a href="https://github.com/01-ai/Yi/blob/main/LICENSE">
32
+ <img src="https://img.shields.io/badge/Code_License-Apache_2.0-lightblue">
33
+ </a>
34
+ </div>
35
+
36
+ <div style="display: inline-block;">
37
+ <a href="https://github.com/01-ai/Yi/blob/main/MODEL_LICENSE_AGREEMENT.txt">
38
+ <img src="https://img.shields.io/badge/Model_License-Yi_License-lightblue">
39
+ </a>
40
+ </div>
41
+
42
+ <div style="display: inline-block;">
43
+ <a href="mailto:[email protected]">
44
+ <img src="https://img.shields.io/badge/✉️[email protected]">
45
+ </a>
46
+ </div>
47
+
48
+ </div>
49
+
50
+ <div align="center">
51
+ <h3 align="center">Building the Next Generation of Open-Source and Bilingual LLMs</h3>
52
+ </div>
53
+
54
+ <p align="center">
55
+ 🤗 <a href="https://huggingface.co/01-ai" target="_blank">Hugging Face</a> • 🤖 <a href="https://www.modelscope.cn/organization/01ai/" target="_blank">ModelScope</a> • ✡️ <a href="https://wisemodel.cn/organization/01.AI" target="_blank">WiseModel</a>
56
+ </p>
57
+
58
+ <p align="center">
59
+ 👋 Join us 💬 <a href="https://github.com/01-ai/Yi/issues/43#issuecomment-1827285245" target="_blank"> WeChat (Chinese) </a>!
60
+ </p>
61
+
62
+
63
+ <!-- DO NOT REMOVE ME -->
64
+
65
+ <hr>
66
+
67
+ <details open>
68
+ <summary></b>📕 Table of Contents</b></summary>
69
+
70
+ - [🟢 What is Yi?](#-what-is-yi)
71
+ - [📌 Introduction](#-introduction)
72
+ - [🎯 Models](#-models)
73
+ - [Chat models](#chat-models)
74
+ - [Base models](#base-models)
75
+ - [Other info](#other-info)
76
+ - [🎉 News](#-news)
77
+ - [🟢 Why Yi?](#-why-yi)
78
+ - [🌎 Ecosystem](#-ecosystem)
79
+ - [💦 Upstream](#-upstream)
80
+ - [🌊 Downstream](#-downstream)
81
+ - [🔗 Serving](#-serving)
82
+ - [⚙️ Quantitation](#️-quantitation)
83
+ - [🛠️ Fine-tuning](#️-fine-tuning)
84
+ - [📌 Benchmarks](#-benchmarks)
85
+ - [📊 Base model performance](#-base-model-performance)
86
+ - [📊 Chat model performance](#-chat-model-performance)
87
+ - [📊 Quantized chat model performance](#-quantized-chat-model-performance)
88
+ - [⛔️ Limitations of chat model](#️-limitations-of-chat-model)
89
+ - [🟢 Who can use Yi?](#-who-can-use-yi)
90
+ - [🟢 How to use Yi?](#-how-to-use-yi)
91
+ - [1. Prepare development environment](#1-prepare-development-environment)
92
+ - [1.1 Docker](#11-docker)
93
+ - [1.2 Local development environment](#12-local-development-environment)
94
+ - [2. Download the model (optional)](#2-download-the-model-optional)
95
+ - [3. Examples](#3-examples)
96
+ - [3.1 Use the chat model](#31-use-the-chat-model)
97
+ - [3.2 Use the base model](#32-use-the-base-model)
98
+ - [3.3 Finetune from the base model](#33-finetune-from-the-base-model)
99
+ - [3.4 Quantization](#34-quantization)
100
+ - [GPT-Q](#gpt-q)
101
+ - [AWQ](#awq)
102
+ - [🟢 Misc.](#-misc)
103
+ - [📡 Disclaimer](#-disclaimer)
104
+ - [🪪 License](#-license)
105
+
106
+ </details>
107
+
108
+ <hr>
109
+
110
+ # 🟢 What is Yi?
111
+
112
+ ## 📌 Introduction
113
+
114
+ - 🤖 The Yi series models are the next generation of open source large language models trained from strach by [01.AI](https://01.ai/).
115
+
116
+ - 🙌 Targeted as a bilingual language model and trained on 3T multilingual corpus, the Yi series models become one of the strongest LLM worldwide, showing promise in language understanding, commonsense reasoning, reading comprehension, and more. For example,
117
+
118
+ - For English language capability, the Yi series models ranked 2nd (just behind GPT-4), outperforming other LLMs (such as LLaMA2-chat-70B, Claude 2, and ChatGPT) on the [AlpacaEval Leaderboard](https://tatsu-lab.github.io/alpaca_eval/) in Dec 2023.
119
+
120
+ - For Chinese language capability, the Yi series models landed in 2nd place (following GPT4), surpassing other LLMs (such as Baidu ERNIE, Qwen, and Baichuan) on the [SuperCLUE](https://www.superclueai.com/) in Oct 2023.
121
+
122
+ - 🙏 (Credits to LLaMA) Thanks to the Transformer and LLaMA open-source communities, as they reducing the efforts required to build from scratch and enabling the utilization of the same tools within the AI ecosystem. If you're interested in Yi's adoption of LLaMA architecture and license usage policy, see [Yi's relation with LLaMA](./docs/yi_relation_llama.md).
123
+
124
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
125
+
126
+ ## 🎯 Models
127
+
128
+ Yi models come in multiple sizes and cater to different use cases. You can also fine-tune Yi models to meet your specific requirements.
129
+
130
+ ### Chat models
131
+
132
+ | Model | Download
133
+ |---|---
134
+ Yi-6B-Chat| • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-6B-Chat) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-6B-Chat/summary)
135
+ Yi-6B-Chat-4bits | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-6B-Chat-4bits) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-6B-Chat-4bits/summary)
136
+ Yi-6B-Chat-8bits | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-6B-Chat-8bits) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-6B-Chat-8bits/summary)
137
+ Yi-34B-Chat | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-34B-Chat) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-34B-Chat/summary)
138
+ Yi-34B-Chat-4bits | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-34B-Chat-4bits) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-34B-Chat-4bits/summary)
139
+ Yi-34B-Chat-8bits | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-34B-Chat-8bits) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-34B-Chat-8bits/summary)
140
+
141
+ <sub><sup> - 4 bits series models are quantized by AWQ. <br> - 8 bits series models are quantized by GPTQ <br> - All quantized models have a low barrier to use since they can be deployed on consumer-grade GPUs (e.g., 3090, 4090).</sup></sub>
142
+
143
+ ### Base models
144
+
145
+ | Model | Download |
146
+ |---|---|
147
+ Yi-6B| • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-6B) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-6B/summary)
148
+ Yi-6B-200K | • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-6B-200K) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-6B-200K/summary)
149
+ Yi-34B| • [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-34B) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-34B/summary)
150
+ Yi-34B-200K|• [🤗 Hugging Face](https://huggingface.co/01-ai/Yi-34B-200K) • [🤖 ModelScope](https://www.modelscope.cn/models/01ai/Yi-34B-200K/summary)
151
+
152
+ <sub><sup> - 200k is roughly equivalent to 400,000 Chinese characters. </sup></sub>
153
+
154
+ ### Other info
155
+
156
+ For chat models and base models:
157
+
158
+ - 6B series models are suitable for personal and academic use.
159
+
160
+ - 34B series models suitable for personal, academic, and commercial (particularly for small and medium-sized enterprises) purposes. It's a cost-effective solution that's affordable and equipped with emergent ability.
161
+
162
+ - The **default context window** is **4k tokens**.
163
+
164
+ - The pretrained tokens are 3T.
165
+
166
+ - The training data are up to June 2023.
167
+
168
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
169
+
170
+ ## 🎉 News
171
+
172
+ <details>
173
+ <summary>🎯 <b>2023/11/23</b>: The chat models are open to public.</summary>
174
+
175
+ This release contains two chat models based on previous released base models, two 8-bits models quantized by GPTQ, two 4-bits models quantized by AWQ.
176
+
177
+ - `Yi-34B-Chat`
178
+ - `Yi-34B-Chat-4bits`
179
+ - `Yi-34B-Chat-8bits`
180
+ - `Yi-6B-Chat`
181
+ - `Yi-6B-Chat-4bits`
182
+ - `Yi-6B-Chat-8bits`
183
+
184
+ You can try some of them interactively at:
185
+
186
+ - [Hugging Face](https://huggingface.co/spaces/01-ai/Yi-34B-Chat)
187
+ - [Replicate](https://replicate.com/01-ai)
188
+ </details>
189
+
190
+ <details>
191
+ <summary>🔔 <b>2023/11/23</b>: The Yi Series Models Community License Agreement is updated to v2.1.</summary>
192
+ </details>
193
+
194
+ <details>
195
+ <summary>🔥 <b>2023/11/08</b>: Invited test of Yi-34B chat model.</summary>
196
+
197
+ Application form:
198
+
199
+ - [English](https://cn.mikecrm.com/l91ODJf)
200
+ - [Chinese](https://cn.mikecrm.com/gnEZjiQ)
201
+
202
+ </details>
203
+
204
+ <details>
205
+ <summary>🎯 <b>2023/11/05</b>: The base model of <code>Yi-6B-200K</code> and <code>Yi-34B-200K</code>.</summary>
206
+
207
+ This release contains two base models with the same parameter sizes of previous
208
+ release, except that the context window is extended to 200K.
209
+
210
+ </details>
211
+
212
+ <details>
213
+ <summary>🎯 <b>2023/11/02</b>: The base model of <code>Yi-6B</code> and <code>Yi-34B</code>.</summary>
214
+
215
+ The first public release contains two bilingual (English/Chinese) base models
216
+ with the parameter sizes of 6B and 34B. Both of them are trained with 4K
217
+ sequence length and can be extended to 32K during inference time.
218
+
219
+ </details>
220
+
221
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
222
+
223
+
224
+ # 🟢 Why Yi?
225
+
226
+ ## 🌎 Ecosystem
227
+
228
+ Yi has a comprehensive ecosystem, offering a range of tools, services, and models to enrich your experiences and maximize productivity.
229
+
230
+ - [💦 Upstream](#-upstream)
231
+ - [🌊 Downstream](#-downstream)
232
+ - [🔗 Serving](#-serving)
233
+ - [⚙️ Quantitation](#️-quantitation)
234
+ - [🛠️ Fine-tuning](#️-fine-tuning)
235
+
236
+ ### 💦 Upstream
237
+
238
+ The Yi series models follow the same model architecture as LLaMA. By choosing Yi, you can leverage existing tools, libraries, and resources within the LLaMA ecosystem, eliminating the need to create new tools and enhancing development efficiency.
239
+
240
+ For example, the Yi series models are saved in the format of the LLaMA model. You can directly use `LLaMAForCausalLM` and `LLaMATokenizer` to load the model. For more information, see [Use the chat model](#31-use-the-chat-model).
241
+
242
+ ```python
243
+ from transformers import AutoModelForCausalLM, AutoTokenizer
244
+
245
+ tokenizer = AutoTokenizer.from_pretrained("01-ai/Yi-34b", use_fast=False)
246
+
247
+ model = AutoModelForCausalLM.from_pretrained("01-ai/Yi-34b", device_map="auto")
248
+ ```
249
+
250
+ ### 🌊 Downstream
251
+
252
+ > 💡 Tip
253
+ >
254
+ > - Feel free to create a PR and share the fantastic work you've built using the Yi series models.
255
+ >
256
+ > - To help others quickly understand your work, it is recommended to use the format of `<model-name>: <model-intro> + <model-highlights>`.
257
+
258
+ #### 🔗 Serving
259
+
260
+ If you want to get up with Yi in a few minutes, you can use the following services built upon Yi.
261
+
262
+ - [Yi-34B-Chat](https://platform.lingyiwanwu.com/) (Yi official beta): you can chat with it. **Note** that currently it's available through a whitelist. Welcome to apply and experience it firsthand!
263
+
264
+ - [Yi-6B-Chat (Replicate)](https://replicate.com/01-ai): you can use this model with more options by setting additional parameters and calling APIs.
265
+
266
+ - [ScaleLLM](https://github.com/vectorch-ai/ScaleLLM#supported-models): you can use this service to run Yi models locally with added flexibility and customization.
267
+
268
+ #### ⚙️ Quantitation
269
+
270
+ If you have limited computational capabilities, you can use Yi's quantized models as follows.
271
+
272
+ These quantized models have reduced precision and but offer increased efficiency, such as faster inference speed and smaller RAM usage.
273
+
274
+ - [TheBloke/Yi-34B-GPTQ](https://huggingface.co/TheBloke/Yi-34B-GPTQ)
275
+ - [TheBloke/Yi-34B-GGUF](https://huggingface.co/TheBloke/Yi-34B-GGUF)
276
+ - [TheBloke/Yi-34B-AWQ](https://huggingface.co/TheBloke/Yi-34B-AWQ)
277
+
278
+ #### 🛠️ Fine-tuning
279
+
280
+ If you're seeking to explore the diverse capabilities within Yi's thriving family, you can delve into Yi's fine-tuned models as below.
281
+
282
+ - [TheBloke Models](https://huggingface.co/TheBloke): this site hosts numerous fine-tuned models derived from various LLMs including Yi.
283
+
284
+ This is not an exhaustive list for Yi, but to name a few sorted on downloads:
285
+ - [TheBloke/dolphin-2_2-yi-34b-AWQ](https://huggingface.co/TheBloke/dolphin-2_2-yi-34b-AWQ)
286
+ - [TheBloke/Yi-34B-Chat-AWQ](https://huggingface.co/TheBloke/Yi-34B-Chat-AWQ)
287
+ - [TheBloke/Yi-34B-Chat-GPTQ](https://huggingface.co/TheBloke/Yi-34B-Chat-GPTQ)
288
+
289
+ - [SUSTech/SUS-Chat-34B](https://huggingface.co/SUSTech/SUS-Chat-34B): this model ranked first among all models below 70B and outperformed the twice larger deepseek-llm-67b-chat. You can check the result on the [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard).
290
+
291
+ - [OrionStarAI/OrionStar-Yi-34B-Chat-Llama](https://huggingface.co/OrionStarAI/OrionStar-Yi-34B-Chat-Llama): this model excelled beyond other models (such as GPT-4, Qwen-14B-Chat, Baichuan2-13B-Chat) in C-Eval and CMMLU evaluations on the [OpenCompass LLM Leaderboard](https://opencompass.org.cn/leaderboard-llm).
292
+
293
+ - [NousResearch/Nous-Capybara-34B](https://huggingface.co/NousResearch/Nous-Capybara-34B): this model is trained with 200K context length and 3 epochs on the Capybara dataset.
294
+
295
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
296
+
297
+
298
+ ## 📌 Benchmarks
299
+
300
+ - [📊 Base model performance](#-base-model-performance)
301
+ - [📊 Chat model performance](#-chat-model-performance)
302
+ - [📊 Quantized chat model performance](#-quantized-chat-model-performance)
303
+ - [⛔️ Limitations of chat model](#️-limitations-of-chat-model)
304
+
305
+ ### 📊 Base model performance
306
+
307
+ | Model | MMLU | CMMLU | C-Eval | GAOKAO | BBH | Common-sense Reasoning | Reading Comprehension | Math & Code |
308
+ | :------------ | :------: | :------: | :------: | :------: | :------: | :--------------------: | :-------------------: | :---------: |
309
+ | | 5-shot | 5-shot | 5-shot | 0-shot | 3-shot@1 | - | - | - |
310
+ | LLaMA2-34B | 62.6 | - | - | - | 44.1 | 69.9 | 68.0 | 26.0 |
311
+ | LLaMA2-70B | 68.9 | 53.3 | - | 49.8 | 51.2 | 71.9 | 69.4 | 36.8 |
312
+ | Baichuan2-13B | 59.2 | 62.0 | 58.1 | 54.3 | 48.8 | 64.3 | 62.4 | 23.0 |
313
+ | Qwen-14B | 66.3 | 71.0 | 72.1 | 62.5 | 53.4 | 73.3 | 72.5 | **39.8** |
314
+ | Skywork-13B | 62.1 | 61.8 | 60.6 | 68.1 | 41.7 | 72.4 | 61.4 | 24.9 |
315
+ | InternLM-20B | 62.1 | 59.0 | 58.8 | 45.5 | 52.5 | 78.3 | - | 30.4 |
316
+ | Aquila-34B | 67.8 | 71.4 | 63.1 | - | - | - | - | - |
317
+ | Falcon-180B | 70.4 | 58.0 | 57.8 | 59.0 | 54.0 | 77.3 | 68.8 | 34.0 |
318
+ | Yi-6B | 63.2 | 75.5 | 72.0 | 72.2 | 42.8 | 72.3 | 68.7 | 19.8 |
319
+ | Yi-6B-200K | 64.0 | 75.3 | 73.5 | 73.9 | 42.0 | 72.0 | 69.1 | 19.0 |
320
+ | **Yi-34B** | **76.3** | **83.7** | 81.4 | 82.8 | **54.3** | **80.1** | 76.4 | 37.1 |
321
+ | Yi-34B-200K | 76.1 | 83.6 | **81.9** | **83.4** | 52.7 | 79.7 | **76.6** | 36.3 |
322
+
323
+ While benchmarking open-source models, we have observed a disparity between the
324
+ results generated by our pipeline and those reported in public sources (e.g.
325
+ OpenCompass). Upon conducting a more in-depth investigation of this difference,
326
+ we have discovered that various models may employ different prompts,
327
+ post-processing strategies, and sampling techniques, potentially resulting in
328
+ significant variations in the outcomes. Our prompt and post-processing strategy
329
+ remains consistent with the original benchmark, and greedy decoding is employed
330
+ during evaluation without any post-processing for the generated content. For
331
+ scores that were not reported by the original authors (including scores reported
332
+ with different settings), we try to get results with our pipeline.
333
+
334
+ To evaluate the model's capability extensively, we adopted the methodology
335
+ outlined in Llama2. Specifically, we included PIQA, SIQA, HellaSwag, WinoGrande,
336
+ ARC, OBQA, and CSQA to assess common sense reasoning. SquAD, QuAC, and BoolQ
337
+ were incorporated to evaluate reading comprehension. CSQA was exclusively tested
338
+ using a 7-shot setup, while all other tests were conducted with a 0-shot
339
+ configuration. Additionally, we introduced GSM8K (8-shot@1), MATH (4-shot@1),
340
+ HumanEval (0-shot@1), and MBPP (3-shot@1) under the category "Math & Code". Due
341
+ to technical constraints, we did not test Falcon-180 on QuAC and OBQA; the score
342
+ is derived by averaging the scores on the remaining tasks. Since the scores for
343
+ these two tasks are generally lower than the average, we believe that
344
+ Falcon-180B's performance was not underestimated.
345
+
346
+ ### 📊 Chat model performance
347
+
348
+ | Model | MMLU | MMLU | CMMLU | CMMLU | C-Eval(val)<sup>*</sup> | C-Eval(val)<sup>*</sup> | Truthful QA | BBH | BBH | GSM8k | GSM8k |
349
+ | ----------------------- | --------- | --------- | --------- | --------- | ----------------------- | ----------------------- | ----------- | --------- | --------- | --------- | --------- |
350
+ | | 0-shot | 5-shot | 0-shot | 5-shot | 0-shot | 5-shot | 0-shot | 0-shot | 3-shot | 0-shot | 4-shot |
351
+ | LLaMA2-13B-Chat | 50.88 | 47.33 | 27.47 | 35.08 | 27.93 | 35.88 | 36.84 | 32.90 | 58.22 | 36.85 | 2.73 |
352
+ | LLaMA2-70B-Chat | 59.42 | 59.86 | 36.10 | 40.99 | 34.99 | 41.31 | 53.95 | 42.36 | 58.53 | 47.08 | 58.68 |
353
+ | Baichuan2-13B-Chat | 55.09 | 50.14 | 58.64 | 59.47 | 56.02 | 54.75 | 48.98 | 38.81 | 47.15 | 45.72 | 23.28 |
354
+ | Qwen-14B-Chat | 63.99 | 64.98 | 67.73 | 70.57 | 66.12 | 70.06 | 52.49 | 49.65 | 54.98 | 59.51 | 61.18 |
355
+ | InternLM-Chat-20B | 55.55 | 57.42 | 53.55 | 53.75 | 51.19 | 53.57 | 51.75 | 42.41 | 36.68 | 15.69 | 43.44 |
356
+ | AquilaChat2-34B v1.2 | 65.15 | 66.70 | 67.51 | 70.02 | **82.99** | **89.38** | **64.33** | 20.12 | 34.28 | 11.52 | 48.45 |
357
+ | Yi-6B-Chat | 58.24 | 60.99 | 69.44 | 74.71 | 68.80 | 74.22 | 50.58 | 39.70 | 47.15 | 38.44 | 44.88 |
358
+ | Yi-6B-Chat-8bits(GPTQ) | 58.29 | 60.96 | 69.21 | 74.69 | 69.17 | 73.85 | 49.85 | 40.35 | 47.26 | 39.42 | 44.88 |
359
+ | Yi-6B-Chat-4bits(AWQ) | 56.78 | 59.89 | 67.70 | 73.29 | 67.53 | 72.29 | 50.29 | 37.74 | 43.62 | 35.71 | 38.36 |
360
+ | Yi-34B-Chat | **67.62** | 73.46 | **79.11** | **81.34** | 77.04 | 78.53 | 62.43 | 51.41 | **71.74** | **71.65** | **75.97** |
361
+ | Yi-34B-Chat-8bits(GPTQ) | 66.24 | **73.69** | 79.05 | 81.23 | 76.82 | 78.97 | 61.84 | **52.08** | 70.97 | 70.74 | 75.74 |
362
+ | Yi-34B-Chat-4bits(AWQ) | 65.77 | 72.42 | 78.21 | 80.50 | 75.71 | 77.27 | 61.84 | 48.30 | 69.39 | 70.51 | 74.00 |
363
+
364
+ We evaluated various benchmarks using both zero-shot and few-shot methods, except for TruthfulQA. Generally, the zero-shot approach is more common in chat models. Our evaluation strategy involves generating responses while following instructions explicitly or implicitly (such as using few-shot examples). We then isolate relevant answers from the generated text. Some models are not well-suited to produce output in the specific format required by instructions in few datasets, which leads to suboptimal results.
365
+
366
+ <strong>*</strong>: C-Eval results are evaluated on the validation datasets
367
+
368
+ ### 📊 Quantized chat model performance
369
+
370
+ We also provide both 4-bit (AWQ) and 8-bit (GPTQ) quantized Yi chat models. Evaluation results on various benchmarks have shown that the quantized models have negligible losses. Additionally, they reduce the memory footprint size. After testing different configurations of prompts and generation lengths, we highly recommend following the guidelines in the memory footprint table below when selecting a device to run our models.
371
+
372
+ | | batch=1 | batch=4 | batch=16 | batch=32 |
373
+ | ----------------------- | ------- | ------- | -------- | -------- |
374
+ | Yi-34B-Chat | 65GiB | 68GiB | 76GiB | >80GiB |
375
+ | Yi-34B-Chat-8bits(GPTQ) | 35GiB | 37GiB | 46GiB | 58GiB |
376
+ | Yi-34B-Chat-4bits(AWQ) | 19GiB | 20GiB | 30GiB | 40GiB |
377
+ | Yi-6B-Chat | 12GiB | 13GiB | 15GiB | 18GiB |
378
+ | Yi-6B-Chat-8bits(GPTQ) | 7GiB | 8GiB | 10GiB | 14GiB |
379
+ | Yi-6B-Chat-4bits(AWQ) | 4GiB | 5GiB | 7GiB | 10GiB |
380
+
381
+ Note: All the numbers in the table represent the minimum recommended memory for running models of the corresponding size.
382
+
383
+ ### ⛔️ Limitations of chat model
384
+
385
+ The released chat model has undergone exclusive training using Supervised Fine-Tuning (SFT). Compared to other standard chat models, our model produces more diverse responses, making it suitable for various downstream tasks, such as creative scenarios. Furthermore, this diversity is expected to enhance the likelihood of generating higher quality responses, which will be advantageous for subsequent Reinforcement Learning (RL) training.
386
+
387
+ However, this higher diversity might amplify certain existing issues, including:
388
+
389
+ - **Hallucination**: This refers to the model generating factually incorrect or nonsensical information. With the model's responses being more varied, there's a higher chance of hallucination that are not based on accurate data or logical reasoning.
390
+ - **Non-determinism in re-generation**: When attempting to regenerate or sample responses, inconsistencies in the outcomes may occur. The increased diversity can lead to varying results even under similar input conditions.
391
+ - **Cumulative Error**: This occurs when errors in the model's responses compound over time. As the model generates more diverse responses, the likelihood of small inaccuracies building up into larger errors increases, especially in complex tasks like extended reasoning, mathematical problem-solving, etc.
392
+
393
+ To achieve more coherent and consistent responses, it is advisable to adjust generation configuration parameters such as`temperature`,`top_p`, or`top_k`. These adjustments can help in the balance between creativity and coherence in the model's outputs.
394
+
395
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
396
+
397
+
398
+ # 🟢 Who can use Yi?
399
+
400
+ Everyone! 🙌 ✅
401
+
402
+ - The Yi series models are free for personal usage, academic purposes, and commercial use. All usage must adhere to the [Yi Series Models Community License Agreement 2.1](https://github.com/01-ai/Yi/blob/main/MODEL_LICENSE_AGREEMENT.txt)
403
+
404
+ - For free commercial use, you only need to [complete this form](https://www.lingyiwanwu.com/yi-license) to get Yi Model Commercial License.
405
+
406
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
407
+
408
+ # 🟢 How to use Yi?
409
+
410
+ [1. Prepare development environment](#1-prepare-development-environment)
411
+ <br>[2. Download the model](#2-download-the-model-optional)
412
+ <br>[3. Examples](#3-examples)
413
+
414
+ ### 1. Prepare development environment
415
+
416
+ #### 1.1 Docker
417
+ The best approach to try the **Yi** series models is through Docker with GPUs. We
418
+ provide the following docker images to help you get started.
419
+
420
+ - `registry.lingyiwanwu.com/ci/01-ai/yi:latest`
421
+ - `ghcr.io/01-ai/yi:latest`
422
+
423
+ Note that the `latest` tag always points to the latest code in the `main`
424
+ branch. To test a stable version, please replace it with a specific
425
+ [tag](https://github.com/01-ai/Yi/tags).
426
+
427
+ #### 1.2 Local development environment
428
+ We use [`conda-lock`](https://github.com/conda/conda-lock) to generate fully reproducible lock files for conda environments. You can refer to [conda-lock.yml](./conda-lock.yml) for the exact versions of the dependencies. Additionally, we utilize [`micromamba`](https://mamba.readthedocs.io/en/latest/user_guide/micromamba.html) for installing these dependencies.
429
+
430
+ To install the dependencies, please follow these steps:
431
+ 1. Install `micromamba` by following the instructions available [here](https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html).
432
+ 2. Execute `micromamba install -y -n yi -f conda-lock.yml` to create a conda environment named `yi` and install the necessary dependencies.
433
+
434
+ ### 2. Download the model (optional)
435
+
436
+ By default, the model weights and tokenizer will be downloaded from
437
+ [Hugging Face](https://huggingface.co/01-ai) automatically in the next step. You
438
+ can also download them manually from the following places:
439
+
440
+ - [ModelScope](https://www.modelscope.cn/organization/01ai/)
441
+ - [WiseModel](https://wisemodel.cn/organization/01.AI)
442
+
443
+ ### 3. Examples
444
+
445
+ #### 3.1 Use the chat model
446
+
447
+ ```python
448
+ from transformers import AutoModelForCausalLM, AutoTokenizer
449
+
450
+ model_path = '01-ai/Yi-34b-Chat'
451
+
452
+ tokenizer = AutoTokenizer.from_pretrained(model_path, use_fast=False)
453
+
454
+ # Since transformers 4.35.0, the GPT-Q/AWQ model can be loaded using AutoModelForCausalLM.
455
+ model = AutoModelForCausalLM.from_pretrained(
456
+ model_path,
457
+ device_map="auto",
458
+ torch_dtype='auto'
459
+ ).eval()
460
+
461
+ # Prompt content: "hi"
462
+ messages = [
463
+ {"role": "user", "content": "hi"}
464
+ ]
465
+
466
+ input_ids = tokenizer.apply_chat_template(conversation=messages, tokenize=True, add_generation_prompt=True, return_tensors='pt')
467
+ output_ids = model.generate(input_ids.to('cuda'))
468
+ response = tokenizer.decode(output_ids[0][input_ids.shape[1]:], skip_special_tokens=True)
469
+
470
+ # Model response: "Hello! How can I assist you today?"
471
+ print(response)
472
+ ```
473
+
474
+
475
+ To construct the prompt template manually, you can refer the `chat_template` field in the `tokenizer_config.json` [file](https://huggingface.co/01-ai/Yi-34B-Chat/blob/main/tokenizer_config.json#L60).
476
+
477
+ ```
478
+ <|im_start|>system
479
+ {system_message}<|im_end|>
480
+ <|im_start|>user
481
+ {prompt}<|im_end|>
482
+ <|im_start|>assistant
483
+ ```
484
+
485
+ #### 3.2 Use the base model
486
+
487
+ ```bash
488
+ python demo/text_generation.py
489
+ ```
490
+
491
+ To reuse the downloaded models in the previous step, you can provide the extra
492
+ `--model` argument:
493
+
494
+ ```bash
495
+ python demo/text_generation.py --model /path/to/model
496
+ ```
497
+
498
+ Or if you'd like to get your hands dirty:
499
+
500
+ ```python
501
+ from transformers import AutoModelForCausalLM, AutoTokenizer
502
+
503
+ model = AutoModelForCausalLM.from_pretrained("01-ai/Yi-34B", device_map="auto", torch_dtype="auto")
504
+ tokenizer = AutoTokenizer.from_pretrained("01-ai/Yi-34B")
505
+ inputs = tokenizer("There's a place where time stands still. A place of breath taking wonder, but also", return_tensors="pt")
506
+ max_length = 256
507
+
508
+ outputs = model.generate(
509
+ inputs.input_ids.cuda(),
510
+ max_length=max_length,
511
+ eos_token_id=tokenizer.eos_token_id,
512
+ do_sample=True,
513
+ repetition_penalty=1.3,
514
+ no_repeat_ngram_size=5,
515
+ temperature=0.7,
516
+ top_k=40,
517
+ top_p=0.8,
518
+ )
519
+ print(tokenizer.decode(outputs[0], skip_special_tokens=True))
520
+ ```
521
+
522
+ <details>
523
+
524
+ <summary>Output</summary>
525
+
526
+ **Prompt**: There's a place where time stands still. A place of breath taking wonder, but also
527
+
528
+ **Generation**: There's a place where time stands still. A place of breath taking wonder, but also of great danger. A place where the very air you breathe could kill you. A place where the only way to survive is to be prepared.
529
+ The place is called the Arctic.
530
+ The Arctic is a vast, frozen wilderness. It is a place of extremes. The temperatures can drop to -40 degrees Celsius. The winds can reach speeds of 100 kilometers per hour. The sun can shine for 24 hours a day, or not at all for weeks on end.
531
+ The Arctic is also a place of great beauty. The ice and snow are a pristine white. The sky is a deep blue. The sunsets are spectacular.
532
+ But the Arctic is also a place of great danger. The ice can be treacherous. The winds can be deadly. The sun can be blinding.
533
+ The Arctic is a place where the only way to survive is to be prepared.
534
+ The Arctic is a place of extremes. The temperatures can drop to -40 degrees Celsius. The winds can reach speeds of 100 kilometers per hour. The sun can shine for 24 hours a day, or not at all for weeks on end.
535
+ The Arctic is a place of great beauty. The ice and snow are a
536
+
537
+ </details>
538
+
539
+ For more advanced usage, please refer to the
540
+ [doc](https://github.com/01-ai/Yi/tree/main/demo).
541
+
542
+ #### 3.3 Finetune from the base model
543
+
544
+ ```bash
545
+ bash finetune/scripts/run_sft_Yi_6b.sh
546
+ ```
547
+
548
+ Once finished, you can compare the finetuned model and the base model with the following command:
549
+
550
+ ```bash
551
+ bash finetune/scripts/run_eval.sh
552
+ ```
553
+
554
+ For more advanced usage like fine-tuning based on your custom data, please refer
555
+ the [doc](https://github.com/01-ai/Yi/tree/main/finetune).
556
+
557
+ #### 3.4 Quantization
558
+
559
+ ##### GPT-Q
560
+ ```bash
561
+ python quantization/gptq/quant_autogptq.py \
562
+ --model /base_model \
563
+ --output_dir /quantized_model \
564
+ --trust_remote_code
565
+ ```
566
+
567
+ Once finished, you can then evaluate the resulting model as follows:
568
+
569
+ ```bash
570
+ python quantization/gptq/eval_quantized_model.py \
571
+ --model /quantized_model \
572
+ --trust_remote_code
573
+ ```
574
+
575
+ For a more detailed explanation, please read the [doc](https://github.com/01-ai/Yi/tree/main/quantization/gptq)
576
+
577
+ ##### AWQ
578
+ ```bash
579
+ python quantization/awq/quant_autoawq.py \
580
+ --model /base_model \
581
+ --output_dir /quantized_model \
582
+ --trust_remote_code
583
+ ```
584
+
585
+ Once finished, you can then evaluate the resulting model as follows:
586
+
587
+ ```bash
588
+ python quantization/awq/eval_quantized_model.py \
589
+ --model /quantized_model \
590
+ --trust_remote_code
591
+ ```
592
+
593
+ For more detailed explanation, please read the [doc](https://github.com/01-ai/Yi/tree/main/quantization/awq)
594
+
595
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
596
+
597
+ # 🟢 Misc.
598
+
599
+ ### 📡 Disclaimer
600
+
601
+ We use data compliance checking algorithms during the training process, to
602
+ ensure the compliance of the trained model to the best of our ability. Due to
603
+ complex data and the diversity of language model usage scenarios, we cannot
604
+ guarantee that the model will generate correct, and reasonable output in all
605
+ scenarios. Please be aware that there is still a risk of the model producing
606
+ problematic outputs. We will not be responsible for any risks and issues
607
+ resulting from misuse, misguidance, illegal usage, and related misinformation,
608
+ as well as any associated data security concerns.
609
+
610
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
611
+
612
+
613
+ ### 🪪 License
614
+
615
+ The source code in this repo is licensed under the [Apache 2.0
616
+ license](https://github.com/01-ai/Yi/blob/main/LICENSE). The Yi series models
617
+ are fully open for academic research and free commercial usage with permission
618
+ via applications. All usage must adhere to the [Yi Series Models Community License Agreement 2.1](https://github.com/01-ai/Yi/blob/main/MODEL_LICENSE_AGREEMENT.txt).
619
+ For free commercial use, you only need to send an email to [get official commercial permission](https://www.lingyiwanwu.com/yi-license).
620
+
621
+ <div align="right"> [ <a href="#building-the-next-generation-of-open-source-and-bilingual-llms">Back to top ⬆️ </a> ] </div>
added_tokens.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "<|endoftext|>": 64001,
3
+ "<|startoftext|>": 64000
4
+ }
all_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.7424986867904663,
4
+ "train_runtime": 1158.4304,
5
+ "train_samples_per_second": 5.179,
6
+ "train_steps_per_second": 0.324
7
+ }
config.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "01-ai/Yi-6B-Chat",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "bos_token_id": 1,
8
+ "eos_token_id": 2,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 4096,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 11008,
13
+ "max_position_embeddings": 4096,
14
+ "model_type": "llama",
15
+ "num_attention_heads": 32,
16
+ "num_hidden_layers": 32,
17
+ "num_key_value_heads": 4,
18
+ "pretraining_tp": 1,
19
+ "rms_norm_eps": 1e-05,
20
+ "rope_scaling": null,
21
+ "rope_theta": 5000000.0,
22
+ "tie_word_embeddings": false,
23
+ "torch_dtype": "float16",
24
+ "transformers_version": "4.34.1",
25
+ "use_cache": false,
26
+ "vocab_size": 64064
27
+ }
generation_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 6,
3
+ "do_sample": true,
4
+ "eos_token_id": 7,
5
+ "max_length": 4096,
6
+ "pad_token_id": 0,
7
+ "temperature": 0.6,
8
+ "top_p": 0.8,
9
+ "transformers_version": "4.34.1"
10
+ }
pytorch_model-00001-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb0aedf02d875d3bfc90b5edfe4553bc0c404db463a63c98f9792f1a25358aff
3
+ size 9943653898
pytorch_model-00002-of-00002.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f1a7ee5191394d48897582b045d99447be3f584345826a57cda18731d7005b17
3
+ size 2991168560
pytorch_model.bin.index.json ADDED
@@ -0,0 +1,298 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 12934717440
4
+ },
5
+ "weight_map": {
6
+ "lm_head.weight": "pytorch_model-00002-of-00002.bin",
7
+ "model.embed_tokens.weight": "pytorch_model-00001-of-00002.bin",
8
+ "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
9
+ "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
10
+ "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
11
+ "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
12
+ "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
13
+ "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
14
+ "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
15
+ "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
16
+ "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
17
+ "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
18
+ "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
19
+ "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
20
+ "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
21
+ "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
22
+ "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
23
+ "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
24
+ "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
25
+ "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
26
+ "model.layers.10.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
27
+ "model.layers.10.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
28
+ "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
29
+ "model.layers.10.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
30
+ "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
31
+ "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
32
+ "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
33
+ "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
34
+ "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
35
+ "model.layers.11.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
36
+ "model.layers.11.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
37
+ "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
38
+ "model.layers.11.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
39
+ "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
40
+ "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
41
+ "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
42
+ "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
43
+ "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
44
+ "model.layers.12.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
45
+ "model.layers.12.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
46
+ "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
47
+ "model.layers.12.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
48
+ "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
49
+ "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
50
+ "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
51
+ "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
52
+ "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
53
+ "model.layers.13.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
54
+ "model.layers.13.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
55
+ "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
56
+ "model.layers.13.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
57
+ "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
58
+ "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
59
+ "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
60
+ "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
61
+ "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
62
+ "model.layers.14.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
63
+ "model.layers.14.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
64
+ "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
65
+ "model.layers.14.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
66
+ "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
67
+ "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
68
+ "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
69
+ "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
70
+ "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
71
+ "model.layers.15.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
72
+ "model.layers.15.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
73
+ "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
74
+ "model.layers.15.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
75
+ "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
76
+ "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
77
+ "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
78
+ "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
79
+ "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
80
+ "model.layers.16.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
81
+ "model.layers.16.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
82
+ "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
83
+ "model.layers.16.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
84
+ "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
85
+ "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
86
+ "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
87
+ "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
88
+ "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
89
+ "model.layers.17.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
90
+ "model.layers.17.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
91
+ "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
92
+ "model.layers.17.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
93
+ "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
94
+ "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
95
+ "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
96
+ "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
97
+ "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
98
+ "model.layers.18.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
99
+ "model.layers.18.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
100
+ "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
101
+ "model.layers.18.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
102
+ "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
103
+ "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
104
+ "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
105
+ "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
106
+ "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
107
+ "model.layers.19.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
108
+ "model.layers.19.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
109
+ "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
110
+ "model.layers.19.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
111
+ "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
112
+ "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
113
+ "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
114
+ "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
115
+ "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
116
+ "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
117
+ "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
118
+ "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
119
+ "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
120
+ "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
121
+ "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
122
+ "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
123
+ "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
124
+ "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
125
+ "model.layers.20.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
126
+ "model.layers.20.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
127
+ "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
128
+ "model.layers.20.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
129
+ "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
130
+ "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
131
+ "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
132
+ "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
133
+ "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
134
+ "model.layers.21.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
135
+ "model.layers.21.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
136
+ "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
137
+ "model.layers.21.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
138
+ "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
139
+ "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
140
+ "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
141
+ "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
142
+ "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
143
+ "model.layers.22.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
144
+ "model.layers.22.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
145
+ "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
146
+ "model.layers.22.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
147
+ "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
148
+ "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
149
+ "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
150
+ "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
151
+ "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
152
+ "model.layers.23.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
153
+ "model.layers.23.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
154
+ "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
155
+ "model.layers.23.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
156
+ "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
157
+ "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
158
+ "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
159
+ "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
160
+ "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
161
+ "model.layers.24.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
162
+ "model.layers.24.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
163
+ "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
164
+ "model.layers.24.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
165
+ "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
166
+ "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
167
+ "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
168
+ "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
169
+ "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
170
+ "model.layers.25.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
171
+ "model.layers.25.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
172
+ "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
173
+ "model.layers.25.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
174
+ "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
175
+ "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
176
+ "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
177
+ "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
178
+ "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
179
+ "model.layers.26.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
180
+ "model.layers.26.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
181
+ "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
182
+ "model.layers.26.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
183
+ "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
184
+ "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
185
+ "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
186
+ "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
187
+ "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
188
+ "model.layers.27.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
189
+ "model.layers.27.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
190
+ "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
191
+ "model.layers.27.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
192
+ "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
193
+ "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
194
+ "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
195
+ "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
196
+ "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
197
+ "model.layers.28.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
198
+ "model.layers.28.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
199
+ "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
200
+ "model.layers.28.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
201
+ "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
202
+ "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
203
+ "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.bin",
204
+ "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
205
+ "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
206
+ "model.layers.29.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
207
+ "model.layers.29.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
208
+ "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
209
+ "model.layers.29.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
210
+ "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
211
+ "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
212
+ "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.bin",
213
+ "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
214
+ "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
215
+ "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
216
+ "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
217
+ "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
218
+ "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
219
+ "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
220
+ "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
221
+ "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
222
+ "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
223
+ "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
224
+ "model.layers.30.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
225
+ "model.layers.30.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
226
+ "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
227
+ "model.layers.30.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
228
+ "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
229
+ "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
230
+ "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.bin",
231
+ "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
232
+ "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
233
+ "model.layers.31.input_layernorm.weight": "pytorch_model-00002-of-00002.bin",
234
+ "model.layers.31.mlp.down_proj.weight": "pytorch_model-00002-of-00002.bin",
235
+ "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00002-of-00002.bin",
236
+ "model.layers.31.mlp.up_proj.weight": "pytorch_model-00002-of-00002.bin",
237
+ "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00002-of-00002.bin",
238
+ "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00002-of-00002.bin",
239
+ "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00002-of-00002.bin",
240
+ "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00002-of-00002.bin",
241
+ "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00002-of-00002.bin",
242
+ "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
243
+ "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
244
+ "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
245
+ "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
246
+ "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
247
+ "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
248
+ "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
249
+ "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
250
+ "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
251
+ "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
252
+ "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
253
+ "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
254
+ "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
255
+ "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
256
+ "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
257
+ "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
258
+ "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
259
+ "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
260
+ "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
261
+ "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
262
+ "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
263
+ "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
264
+ "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
265
+ "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
266
+ "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
267
+ "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
268
+ "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
269
+ "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
270
+ "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
271
+ "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
272
+ "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
273
+ "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
274
+ "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
275
+ "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
276
+ "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
277
+ "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
278
+ "model.layers.8.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
279
+ "model.layers.8.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
280
+ "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
281
+ "model.layers.8.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
282
+ "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
283
+ "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
284
+ "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
285
+ "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
286
+ "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
287
+ "model.layers.9.input_layernorm.weight": "pytorch_model-00001-of-00002.bin",
288
+ "model.layers.9.mlp.down_proj.weight": "pytorch_model-00001-of-00002.bin",
289
+ "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00001-of-00002.bin",
290
+ "model.layers.9.mlp.up_proj.weight": "pytorch_model-00001-of-00002.bin",
291
+ "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00001-of-00002.bin",
292
+ "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00001-of-00002.bin",
293
+ "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00001-of-00002.bin",
294
+ "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00001-of-00002.bin",
295
+ "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00001-of-00002.bin",
296
+ "model.norm.weight": "pytorch_model-00002-of-00002.bin"
297
+ }
298
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|im_sep|>"
6
+ ],
7
+ "bos_token": {
8
+ "content": "<|startoftext|>",
9
+ "lstrip": false,
10
+ "normalized": true,
11
+ "rstrip": false,
12
+ "single_word": false
13
+ },
14
+ "eos_token": {
15
+ "content": "<|endoftext|>",
16
+ "lstrip": false,
17
+ "normalized": true,
18
+ "rstrip": false,
19
+ "single_word": false
20
+ },
21
+ "pad_token": {
22
+ "content": "<unk>",
23
+ "lstrip": false,
24
+ "normalized": true,
25
+ "rstrip": false,
26
+ "single_word": false
27
+ },
28
+ "unk_token": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false
34
+ }
35
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:386c49cf943d71aa110361135338c50e38beeff0a66593480421f37b319e1a39
3
+ size 1033105
tokenizer_config.json ADDED
@@ -0,0 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<unk>",
7
+ "lstrip": false,
8
+ "normalized": true,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<s>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "</s>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "6": {
30
+ "content": "<|im_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "7": {
38
+ "content": "<|im_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "8": {
46
+ "content": "<|im_sep|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "64000": {
54
+ "content": "<|startoftext|>",
55
+ "lstrip": false,
56
+ "normalized": true,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "64001": {
62
+ "content": "<|endoftext|>",
63
+ "lstrip": false,
64
+ "normalized": true,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ }
69
+ },
70
+ "additional_special_tokens": [
71
+ "<|im_start|>",
72
+ "<|im_end|>",
73
+ "<|im_sep|>"
74
+ ],
75
+ "bos_token": "<|startoftext|>",
76
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
77
+ "clean_up_tokenization_spaces": false,
78
+ "eos_token": "<|endoftext|>",
79
+ "legacy": true,
80
+ "model_max_length": 4096,
81
+ "pad_token": "<unk>",
82
+ "padding_side": "right",
83
+ "sp_model_kwargs": {},
84
+ "spaces_between_special_tokens": false,
85
+ "split_special_tokens": false,
86
+ "tokenizer_class": "LlamaTokenizer",
87
+ "unk_token": "<unk>",
88
+ "use_default_system_prompt": true
89
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.7424986867904663,
4
+ "train_runtime": 1158.4304,
5
+ "train_samples_per_second": 5.179,
6
+ "train_steps_per_second": 0.324
7
+ }
trainer_log.jsonl ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"current_steps": 5, "total_steps": 375, "loss": 1.446, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.997807075247146e-05, "epoch": 0.04, "percentage": 1.33, "elapsed_time": "0:00:16", "remaining_time": "0:19:48"}
2
+ {"current_steps": 10, "total_steps": 375, "loss": 1.2288, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.991232148123761e-05, "epoch": 0.08, "percentage": 2.67, "elapsed_time": "0:00:29", "remaining_time": "0:18:06"}
3
+ {"current_steps": 15, "total_steps": 375, "loss": 1.2506, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.980286753286195e-05, "epoch": 0.12, "percentage": 4.0, "elapsed_time": "0:00:43", "remaining_time": "0:17:23"}
4
+ {"current_steps": 20, "total_steps": 375, "loss": 1.1189, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.964990092676263e-05, "epoch": 0.16, "percentage": 5.33, "elapsed_time": "0:00:58", "remaining_time": "0:17:18"}
5
+ {"current_steps": 25, "total_steps": 375, "loss": 1.104, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.9453690018345144e-05, "epoch": 0.2, "percentage": 6.67, "elapsed_time": "0:01:13", "remaining_time": "0:17:02"}
6
+ {"current_steps": 30, "total_steps": 375, "loss": 1.2144, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.9214579028215776e-05, "epoch": 0.24, "percentage": 8.0, "elapsed_time": "0:01:28", "remaining_time": "0:16:54"}
7
+ {"current_steps": 35, "total_steps": 375, "loss": 1.0983, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.893298743830168e-05, "epoch": 0.28, "percentage": 9.33, "elapsed_time": "0:01:42", "remaining_time": "0:16:31"}
8
+ {"current_steps": 40, "total_steps": 375, "loss": 1.1491, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.860940925593703e-05, "epoch": 0.32, "percentage": 10.67, "elapsed_time": "0:01:56", "remaining_time": "0:16:17"}
9
+ {"current_steps": 45, "total_steps": 375, "loss": 1.1163, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.8244412147206284e-05, "epoch": 0.36, "percentage": 12.0, "elapsed_time": "0:02:11", "remaining_time": "0:16:03"}
10
+ {"current_steps": 50, "total_steps": 375, "loss": 1.0851, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.783863644106502e-05, "epoch": 0.4, "percentage": 13.33, "elapsed_time": "0:02:26", "remaining_time": "0:15:52"}
11
+ {"current_steps": 55, "total_steps": 375, "loss": 1.1639, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.7392794005985326e-05, "epoch": 0.44, "percentage": 14.67, "elapsed_time": "0:02:40", "remaining_time": "0:15:35"}
12
+ {"current_steps": 60, "total_steps": 375, "loss": 1.1499, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.690766700109659e-05, "epoch": 0.48, "percentage": 16.0, "elapsed_time": "0:02:54", "remaining_time": "0:15:14"}
13
+ {"current_steps": 65, "total_steps": 375, "loss": 1.0992, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.638410650401267e-05, "epoch": 0.52, "percentage": 17.33, "elapsed_time": "0:03:08", "remaining_time": "0:14:58"}
14
+ {"current_steps": 70, "total_steps": 375, "loss": 1.1705, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.5823031017752485e-05, "epoch": 0.56, "percentage": 18.67, "elapsed_time": "0:03:21", "remaining_time": "0:14:36"}
15
+ {"current_steps": 75, "total_steps": 375, "loss": 1.0919, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.522542485937369e-05, "epoch": 0.6, "percentage": 20.0, "elapsed_time": "0:03:33", "remaining_time": "0:14:14"}
16
+ {"current_steps": 80, "total_steps": 375, "loss": 1.1709, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.4592336433146e-05, "epoch": 0.64, "percentage": 21.33, "elapsed_time": "0:03:48", "remaining_time": "0:14:00"}
17
+ {"current_steps": 85, "total_steps": 375, "loss": 1.2184, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.3924876391293915e-05, "epoch": 0.68, "percentage": 22.67, "elapsed_time": "0:04:00", "remaining_time": "0:13:40"}
18
+ {"current_steps": 90, "total_steps": 375, "loss": 1.1752, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.3224215685535294e-05, "epoch": 0.72, "percentage": 24.0, "elapsed_time": "0:04:14", "remaining_time": "0:13:24"}
19
+ {"current_steps": 95, "total_steps": 375, "loss": 1.1208, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.249158351283414e-05, "epoch": 0.76, "percentage": 25.33, "elapsed_time": "0:04:27", "remaining_time": "0:13:09"}
20
+ {"current_steps": 100, "total_steps": 375, "loss": 1.0406, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.172826515897146e-05, "epoch": 0.8, "percentage": 26.67, "elapsed_time": "0:04:43", "remaining_time": "0:12:59"}
21
+ {"current_steps": 105, "total_steps": 375, "loss": 1.1003, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.093559974371725e-05, "epoch": 0.84, "percentage": 28.0, "elapsed_time": "0:05:23", "remaining_time": "0:13:50"}
22
+ {"current_steps": 110, "total_steps": 375, "loss": 1.0862, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.011497787155938e-05, "epoch": 0.88, "percentage": 29.33, "elapsed_time": "0:05:39", "remaining_time": "0:13:37"}
23
+ {"current_steps": 115, "total_steps": 375, "loss": 1.0911, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.92678391921108e-05, "epoch": 0.92, "percentage": 30.67, "elapsed_time": "0:05:51", "remaining_time": "0:13:13"}
24
+ {"current_steps": 120, "total_steps": 375, "loss": 1.2021, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.8395669874474915e-05, "epoch": 0.96, "percentage": 32.0, "elapsed_time": "0:06:05", "remaining_time": "0:12:56"}
25
+ {"current_steps": 125, "total_steps": 375, "loss": 1.1072, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.7500000000000003e-05, "epoch": 1.0, "percentage": 33.33, "elapsed_time": "0:06:18", "remaining_time": "0:12:36"}
26
+ {"current_steps": 130, "total_steps": 375, "loss": 0.7294, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.6582400877996546e-05, "epoch": 1.04, "percentage": 34.67, "elapsed_time": "0:06:32", "remaining_time": "0:12:18"}
27
+ {"current_steps": 135, "total_steps": 375, "loss": 0.7575, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.564448228912682e-05, "epoch": 1.08, "percentage": 36.0, "elapsed_time": "0:06:47", "remaining_time": "0:12:04"}
28
+ {"current_steps": 140, "total_steps": 375, "loss": 0.7256, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.4687889661302576e-05, "epoch": 1.12, "percentage": 37.33, "elapsed_time": "0:07:01", "remaining_time": "0:11:47"}
29
+ {"current_steps": 145, "total_steps": 375, "loss": 0.6674, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.3714301183045385e-05, "epoch": 1.16, "percentage": 38.67, "elapsed_time": "0:07:14", "remaining_time": "0:11:29"}
30
+ {"current_steps": 150, "total_steps": 375, "loss": 0.6673, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.272542485937369e-05, "epoch": 1.2, "percentage": 40.0, "elapsed_time": "0:07:29", "remaining_time": "0:11:13"}
31
+ {"current_steps": 155, "total_steps": 375, "loss": 0.6968, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.172299551538164e-05, "epoch": 1.24, "percentage": 41.33, "elapsed_time": "0:07:45", "remaining_time": "0:11:00"}
32
+ {"current_steps": 160, "total_steps": 375, "loss": 0.6963, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.0708771752766394e-05, "epoch": 1.28, "percentage": 42.67, "elapsed_time": "0:07:57", "remaining_time": "0:10:41"}
33
+ {"current_steps": 165, "total_steps": 375, "loss": 0.6005, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.9684532864643122e-05, "epoch": 1.32, "percentage": 44.0, "elapsed_time": "0:08:11", "remaining_time": "0:10:25"}
34
+ {"current_steps": 170, "total_steps": 375, "loss": 0.6712, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.8652075714060295e-05, "epoch": 1.36, "percentage": 45.33, "elapsed_time": "0:08:26", "remaining_time": "0:10:10"}
35
+ {"current_steps": 175, "total_steps": 375, "loss": 0.6859, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.761321158169134e-05, "epoch": 1.4, "percentage": 46.67, "elapsed_time": "0:08:42", "remaining_time": "0:09:57"}
36
+ {"current_steps": 180, "total_steps": 375, "loss": 0.6128, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.656976298823284e-05, "epoch": 1.44, "percentage": 48.0, "elapsed_time": "0:08:56", "remaining_time": "0:09:40"}
37
+ {"current_steps": 185, "total_steps": 375, "loss": 0.6395, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.5523560497083926e-05, "epoch": 1.48, "percentage": 49.33, "elapsed_time": "0:09:10", "remaining_time": "0:09:25"}
38
+ {"current_steps": 190, "total_steps": 375, "loss": 0.7517, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.447643950291608e-05, "epoch": 1.52, "percentage": 50.67, "elapsed_time": "0:09:26", "remaining_time": "0:09:11"}
39
+ {"current_steps": 195, "total_steps": 375, "loss": 0.6147, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.3430237011767167e-05, "epoch": 1.56, "percentage": 52.0, "elapsed_time": "0:09:41", "remaining_time": "0:08:56"}
40
+ {"current_steps": 200, "total_steps": 375, "loss": 0.7306, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.238678841830867e-05, "epoch": 1.6, "percentage": 53.33, "elapsed_time": "0:09:53", "remaining_time": "0:08:38"}
41
+ {"current_steps": 205, "total_steps": 375, "loss": 0.7235, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.1347924285939714e-05, "epoch": 1.64, "percentage": 54.67, "elapsed_time": "0:10:33", "remaining_time": "0:08:45"}
42
+ {"current_steps": 210, "total_steps": 375, "loss": 0.595, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.031546713535688e-05, "epoch": 1.68, "percentage": 56.0, "elapsed_time": "0:10:46", "remaining_time": "0:08:28"}
43
+ {"current_steps": 215, "total_steps": 375, "loss": 0.6347, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.9291228247233605e-05, "epoch": 1.72, "percentage": 57.33, "elapsed_time": "0:11:01", "remaining_time": "0:08:11"}
44
+ {"current_steps": 220, "total_steps": 375, "loss": 0.6938, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.827700448461836e-05, "epoch": 1.76, "percentage": 58.67, "elapsed_time": "0:11:16", "remaining_time": "0:07:56"}
45
+ {"current_steps": 225, "total_steps": 375, "loss": 0.6408, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.7274575140626318e-05, "epoch": 1.8, "percentage": 60.0, "elapsed_time": "0:11:31", "remaining_time": "0:07:40"}
46
+ {"current_steps": 230, "total_steps": 375, "loss": 0.6296, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.6285698816954624e-05, "epoch": 1.84, "percentage": 61.33, "elapsed_time": "0:11:44", "remaining_time": "0:07:24"}
47
+ {"current_steps": 235, "total_steps": 375, "loss": 0.6688, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.5312110338697426e-05, "epoch": 1.88, "percentage": 62.67, "elapsed_time": "0:11:58", "remaining_time": "0:07:08"}
48
+ {"current_steps": 240, "total_steps": 375, "loss": 0.661, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.4355517710873184e-05, "epoch": 1.92, "percentage": 64.0, "elapsed_time": "0:12:11", "remaining_time": "0:06:51"}
49
+ {"current_steps": 245, "total_steps": 375, "loss": 0.6846, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3417599122003464e-05, "epoch": 1.96, "percentage": 65.33, "elapsed_time": "0:12:25", "remaining_time": "0:06:35"}
50
+ {"current_steps": 250, "total_steps": 375, "loss": 0.6501, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.2500000000000006e-05, "epoch": 2.0, "percentage": 66.67, "elapsed_time": "0:12:38", "remaining_time": "0:06:19"}
51
+ {"current_steps": 255, "total_steps": 375, "loss": 0.5096, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.1604330125525079e-05, "epoch": 2.04, "percentage": 68.0, "elapsed_time": "0:12:52", "remaining_time": "0:06:03"}
52
+ {"current_steps": 260, "total_steps": 375, "loss": 0.4049, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.0732160807889211e-05, "epoch": 2.08, "percentage": 69.33, "elapsed_time": "0:13:06", "remaining_time": "0:05:47"}
53
+ {"current_steps": 265, "total_steps": 375, "loss": 0.4556, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.88502212844063e-06, "epoch": 2.12, "percentage": 70.67, "elapsed_time": "0:13:19", "remaining_time": "0:05:31"}
54
+ {"current_steps": 270, "total_steps": 375, "loss": 0.4205, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.064400256282757e-06, "epoch": 2.16, "percentage": 72.0, "elapsed_time": "0:13:35", "remaining_time": "0:05:17"}
55
+ {"current_steps": 275, "total_steps": 375, "loss": 0.3667, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.271734841028553e-06, "epoch": 2.2, "percentage": 73.33, "elapsed_time": "0:13:48", "remaining_time": "0:05:01"}
56
+ {"current_steps": 280, "total_steps": 375, "loss": 0.3817, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.508416487165862e-06, "epoch": 2.24, "percentage": 74.67, "elapsed_time": "0:14:02", "remaining_time": "0:04:45"}
57
+ {"current_steps": 285, "total_steps": 375, "loss": 0.4369, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.775784314464717e-06, "epoch": 2.28, "percentage": 76.0, "elapsed_time": "0:14:17", "remaining_time": "0:04:30"}
58
+ {"current_steps": 290, "total_steps": 375, "loss": 0.3672, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.075123608706093e-06, "epoch": 2.32, "percentage": 77.33, "elapsed_time": "0:14:31", "remaining_time": "0:04:15"}
59
+ {"current_steps": 295, "total_steps": 375, "loss": 0.4072, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.4076635668540075e-06, "epoch": 2.36, "percentage": 78.67, "elapsed_time": "0:14:46", "remaining_time": "0:04:00"}
60
+ {"current_steps": 300, "total_steps": 375, "loss": 0.416, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.7745751406263165e-06, "epoch": 2.4, "percentage": 80.0, "elapsed_time": "0:15:01", "remaining_time": "0:03:45"}
61
+ {"current_steps": 305, "total_steps": 375, "loss": 0.3504, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.176968982247514e-06, "epoch": 2.44, "percentage": 81.33, "elapsed_time": "0:15:42", "remaining_time": "0:03:36"}
62
+ {"current_steps": 310, "total_steps": 375, "loss": 0.4019, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.6158934959873353e-06, "epoch": 2.48, "percentage": 82.67, "elapsed_time": "0:15:56", "remaining_time": "0:03:20"}
63
+ {"current_steps": 315, "total_steps": 375, "loss": 0.4057, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.092332998903416e-06, "epoch": 2.52, "percentage": 84.0, "elapsed_time": "0:16:09", "remaining_time": "0:03:04"}
64
+ {"current_steps": 320, "total_steps": 375, "loss": 0.3894, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.6072059940146775e-06, "epoch": 2.56, "percentage": 85.33, "elapsed_time": "0:16:23", "remaining_time": "0:02:48"}
65
+ {"current_steps": 325, "total_steps": 375, "loss": 0.3795, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.1613635589349756e-06, "epoch": 2.6, "percentage": 86.67, "elapsed_time": "0:16:38", "remaining_time": "0:02:33"}
66
+ {"current_steps": 330, "total_steps": 375, "loss": 0.3934, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.7555878527937164e-06, "epoch": 2.64, "percentage": 88.0, "elapsed_time": "0:16:53", "remaining_time": "0:02:18"}
67
+ {"current_steps": 335, "total_steps": 375, "loss": 0.423, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3905907440629752e-06, "epoch": 2.68, "percentage": 89.33, "elapsed_time": "0:17:06", "remaining_time": "0:02:02"}
68
+ {"current_steps": 340, "total_steps": 375, "loss": 0.4121, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.067012561698319e-06, "epoch": 2.72, "percentage": 90.67, "elapsed_time": "0:17:20", "remaining_time": "0:01:47"}
69
+ {"current_steps": 345, "total_steps": 375, "loss": 0.4026, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.854209717842231e-07, "epoch": 2.76, "percentage": 92.0, "elapsed_time": "0:17:36", "remaining_time": "0:01:31"}
70
+ {"current_steps": 350, "total_steps": 375, "loss": 0.4056, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.463099816548579e-07, "epoch": 2.8, "percentage": 93.33, "elapsed_time": "0:17:48", "remaining_time": "0:01:16"}
71
+ {"current_steps": 355, "total_steps": 375, "loss": 0.3625, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.5009907323737825e-07, "epoch": 2.84, "percentage": 94.67, "elapsed_time": "0:18:01", "remaining_time": "0:01:00"}
72
+ {"current_steps": 360, "total_steps": 375, "loss": 0.4279, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.9713246713805588e-07, "epoch": 2.88, "percentage": 96.0, "elapsed_time": "0:18:16", "remaining_time": "0:00:45"}
73
+ {"current_steps": 365, "total_steps": 375, "loss": 0.3752, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.767851876239074e-08, "epoch": 2.92, "percentage": 97.33, "elapsed_time": "0:18:31", "remaining_time": "0:00:30"}
74
+ {"current_steps": 370, "total_steps": 375, "loss": 0.3992, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.192924752854042e-08, "epoch": 2.96, "percentage": 98.67, "elapsed_time": "0:18:45", "remaining_time": "0:00:15"}
75
+ {"current_steps": 375, "total_steps": 375, "loss": 0.3638, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 0.0, "epoch": 3.0, "percentage": 100.0, "elapsed_time": "0:18:58", "remaining_time": "0:00:00"}
76
+ {"current_steps": 375, "total_steps": 375, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 3.0, "percentage": 100.0, "elapsed_time": "0:18:58", "remaining_time": "0:00:00"}
trainer_state.json ADDED
@@ -0,0 +1,478 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 375,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04,
13
+ "learning_rate": 4.997807075247146e-05,
14
+ "loss": 1.446,
15
+ "step": 5
16
+ },
17
+ {
18
+ "epoch": 0.08,
19
+ "learning_rate": 4.991232148123761e-05,
20
+ "loss": 1.2288,
21
+ "step": 10
22
+ },
23
+ {
24
+ "epoch": 0.12,
25
+ "learning_rate": 4.980286753286195e-05,
26
+ "loss": 1.2506,
27
+ "step": 15
28
+ },
29
+ {
30
+ "epoch": 0.16,
31
+ "learning_rate": 4.964990092676263e-05,
32
+ "loss": 1.1189,
33
+ "step": 20
34
+ },
35
+ {
36
+ "epoch": 0.2,
37
+ "learning_rate": 4.9453690018345144e-05,
38
+ "loss": 1.104,
39
+ "step": 25
40
+ },
41
+ {
42
+ "epoch": 0.24,
43
+ "learning_rate": 4.9214579028215776e-05,
44
+ "loss": 1.2144,
45
+ "step": 30
46
+ },
47
+ {
48
+ "epoch": 0.28,
49
+ "learning_rate": 4.893298743830168e-05,
50
+ "loss": 1.0983,
51
+ "step": 35
52
+ },
53
+ {
54
+ "epoch": 0.32,
55
+ "learning_rate": 4.860940925593703e-05,
56
+ "loss": 1.1491,
57
+ "step": 40
58
+ },
59
+ {
60
+ "epoch": 0.36,
61
+ "learning_rate": 4.8244412147206284e-05,
62
+ "loss": 1.1163,
63
+ "step": 45
64
+ },
65
+ {
66
+ "epoch": 0.4,
67
+ "learning_rate": 4.783863644106502e-05,
68
+ "loss": 1.0851,
69
+ "step": 50
70
+ },
71
+ {
72
+ "epoch": 0.44,
73
+ "learning_rate": 4.7392794005985326e-05,
74
+ "loss": 1.1639,
75
+ "step": 55
76
+ },
77
+ {
78
+ "epoch": 0.48,
79
+ "learning_rate": 4.690766700109659e-05,
80
+ "loss": 1.1499,
81
+ "step": 60
82
+ },
83
+ {
84
+ "epoch": 0.52,
85
+ "learning_rate": 4.638410650401267e-05,
86
+ "loss": 1.0992,
87
+ "step": 65
88
+ },
89
+ {
90
+ "epoch": 0.56,
91
+ "learning_rate": 4.5823031017752485e-05,
92
+ "loss": 1.1705,
93
+ "step": 70
94
+ },
95
+ {
96
+ "epoch": 0.6,
97
+ "learning_rate": 4.522542485937369e-05,
98
+ "loss": 1.0919,
99
+ "step": 75
100
+ },
101
+ {
102
+ "epoch": 0.64,
103
+ "learning_rate": 4.4592336433146e-05,
104
+ "loss": 1.1709,
105
+ "step": 80
106
+ },
107
+ {
108
+ "epoch": 0.68,
109
+ "learning_rate": 4.3924876391293915e-05,
110
+ "loss": 1.2184,
111
+ "step": 85
112
+ },
113
+ {
114
+ "epoch": 0.72,
115
+ "learning_rate": 4.3224215685535294e-05,
116
+ "loss": 1.1752,
117
+ "step": 90
118
+ },
119
+ {
120
+ "epoch": 0.76,
121
+ "learning_rate": 4.249158351283414e-05,
122
+ "loss": 1.1208,
123
+ "step": 95
124
+ },
125
+ {
126
+ "epoch": 0.8,
127
+ "learning_rate": 4.172826515897146e-05,
128
+ "loss": 1.0406,
129
+ "step": 100
130
+ },
131
+ {
132
+ "epoch": 0.84,
133
+ "learning_rate": 4.093559974371725e-05,
134
+ "loss": 1.1003,
135
+ "step": 105
136
+ },
137
+ {
138
+ "epoch": 0.88,
139
+ "learning_rate": 4.011497787155938e-05,
140
+ "loss": 1.0862,
141
+ "step": 110
142
+ },
143
+ {
144
+ "epoch": 0.92,
145
+ "learning_rate": 3.92678391921108e-05,
146
+ "loss": 1.0911,
147
+ "step": 115
148
+ },
149
+ {
150
+ "epoch": 0.96,
151
+ "learning_rate": 3.8395669874474915e-05,
152
+ "loss": 1.2021,
153
+ "step": 120
154
+ },
155
+ {
156
+ "epoch": 1.0,
157
+ "learning_rate": 3.7500000000000003e-05,
158
+ "loss": 1.1072,
159
+ "step": 125
160
+ },
161
+ {
162
+ "epoch": 1.04,
163
+ "learning_rate": 3.6582400877996546e-05,
164
+ "loss": 0.7294,
165
+ "step": 130
166
+ },
167
+ {
168
+ "epoch": 1.08,
169
+ "learning_rate": 3.564448228912682e-05,
170
+ "loss": 0.7575,
171
+ "step": 135
172
+ },
173
+ {
174
+ "epoch": 1.12,
175
+ "learning_rate": 3.4687889661302576e-05,
176
+ "loss": 0.7256,
177
+ "step": 140
178
+ },
179
+ {
180
+ "epoch": 1.16,
181
+ "learning_rate": 3.3714301183045385e-05,
182
+ "loss": 0.6674,
183
+ "step": 145
184
+ },
185
+ {
186
+ "epoch": 1.2,
187
+ "learning_rate": 3.272542485937369e-05,
188
+ "loss": 0.6673,
189
+ "step": 150
190
+ },
191
+ {
192
+ "epoch": 1.24,
193
+ "learning_rate": 3.172299551538164e-05,
194
+ "loss": 0.6968,
195
+ "step": 155
196
+ },
197
+ {
198
+ "epoch": 1.28,
199
+ "learning_rate": 3.0708771752766394e-05,
200
+ "loss": 0.6963,
201
+ "step": 160
202
+ },
203
+ {
204
+ "epoch": 1.32,
205
+ "learning_rate": 2.9684532864643122e-05,
206
+ "loss": 0.6005,
207
+ "step": 165
208
+ },
209
+ {
210
+ "epoch": 1.36,
211
+ "learning_rate": 2.8652075714060295e-05,
212
+ "loss": 0.6712,
213
+ "step": 170
214
+ },
215
+ {
216
+ "epoch": 1.4,
217
+ "learning_rate": 2.761321158169134e-05,
218
+ "loss": 0.6859,
219
+ "step": 175
220
+ },
221
+ {
222
+ "epoch": 1.44,
223
+ "learning_rate": 2.656976298823284e-05,
224
+ "loss": 0.6128,
225
+ "step": 180
226
+ },
227
+ {
228
+ "epoch": 1.48,
229
+ "learning_rate": 2.5523560497083926e-05,
230
+ "loss": 0.6395,
231
+ "step": 185
232
+ },
233
+ {
234
+ "epoch": 1.52,
235
+ "learning_rate": 2.447643950291608e-05,
236
+ "loss": 0.7517,
237
+ "step": 190
238
+ },
239
+ {
240
+ "epoch": 1.56,
241
+ "learning_rate": 2.3430237011767167e-05,
242
+ "loss": 0.6147,
243
+ "step": 195
244
+ },
245
+ {
246
+ "epoch": 1.6,
247
+ "learning_rate": 2.238678841830867e-05,
248
+ "loss": 0.7306,
249
+ "step": 200
250
+ },
251
+ {
252
+ "epoch": 1.64,
253
+ "learning_rate": 2.1347924285939714e-05,
254
+ "loss": 0.7235,
255
+ "step": 205
256
+ },
257
+ {
258
+ "epoch": 1.68,
259
+ "learning_rate": 2.031546713535688e-05,
260
+ "loss": 0.595,
261
+ "step": 210
262
+ },
263
+ {
264
+ "epoch": 1.72,
265
+ "learning_rate": 1.9291228247233605e-05,
266
+ "loss": 0.6347,
267
+ "step": 215
268
+ },
269
+ {
270
+ "epoch": 1.76,
271
+ "learning_rate": 1.827700448461836e-05,
272
+ "loss": 0.6938,
273
+ "step": 220
274
+ },
275
+ {
276
+ "epoch": 1.8,
277
+ "learning_rate": 1.7274575140626318e-05,
278
+ "loss": 0.6408,
279
+ "step": 225
280
+ },
281
+ {
282
+ "epoch": 1.84,
283
+ "learning_rate": 1.6285698816954624e-05,
284
+ "loss": 0.6296,
285
+ "step": 230
286
+ },
287
+ {
288
+ "epoch": 1.88,
289
+ "learning_rate": 1.5312110338697426e-05,
290
+ "loss": 0.6688,
291
+ "step": 235
292
+ },
293
+ {
294
+ "epoch": 1.92,
295
+ "learning_rate": 1.4355517710873184e-05,
296
+ "loss": 0.661,
297
+ "step": 240
298
+ },
299
+ {
300
+ "epoch": 1.96,
301
+ "learning_rate": 1.3417599122003464e-05,
302
+ "loss": 0.6846,
303
+ "step": 245
304
+ },
305
+ {
306
+ "epoch": 2.0,
307
+ "learning_rate": 1.2500000000000006e-05,
308
+ "loss": 0.6501,
309
+ "step": 250
310
+ },
311
+ {
312
+ "epoch": 2.04,
313
+ "learning_rate": 1.1604330125525079e-05,
314
+ "loss": 0.5096,
315
+ "step": 255
316
+ },
317
+ {
318
+ "epoch": 2.08,
319
+ "learning_rate": 1.0732160807889211e-05,
320
+ "loss": 0.4049,
321
+ "step": 260
322
+ },
323
+ {
324
+ "epoch": 2.12,
325
+ "learning_rate": 9.88502212844063e-06,
326
+ "loss": 0.4556,
327
+ "step": 265
328
+ },
329
+ {
330
+ "epoch": 2.16,
331
+ "learning_rate": 9.064400256282757e-06,
332
+ "loss": 0.4205,
333
+ "step": 270
334
+ },
335
+ {
336
+ "epoch": 2.2,
337
+ "learning_rate": 8.271734841028553e-06,
338
+ "loss": 0.3667,
339
+ "step": 275
340
+ },
341
+ {
342
+ "epoch": 2.24,
343
+ "learning_rate": 7.508416487165862e-06,
344
+ "loss": 0.3817,
345
+ "step": 280
346
+ },
347
+ {
348
+ "epoch": 2.28,
349
+ "learning_rate": 6.775784314464717e-06,
350
+ "loss": 0.4369,
351
+ "step": 285
352
+ },
353
+ {
354
+ "epoch": 2.32,
355
+ "learning_rate": 6.075123608706093e-06,
356
+ "loss": 0.3672,
357
+ "step": 290
358
+ },
359
+ {
360
+ "epoch": 2.36,
361
+ "learning_rate": 5.4076635668540075e-06,
362
+ "loss": 0.4072,
363
+ "step": 295
364
+ },
365
+ {
366
+ "epoch": 2.4,
367
+ "learning_rate": 4.7745751406263165e-06,
368
+ "loss": 0.416,
369
+ "step": 300
370
+ },
371
+ {
372
+ "epoch": 2.44,
373
+ "learning_rate": 4.176968982247514e-06,
374
+ "loss": 0.3504,
375
+ "step": 305
376
+ },
377
+ {
378
+ "epoch": 2.48,
379
+ "learning_rate": 3.6158934959873353e-06,
380
+ "loss": 0.4019,
381
+ "step": 310
382
+ },
383
+ {
384
+ "epoch": 2.52,
385
+ "learning_rate": 3.092332998903416e-06,
386
+ "loss": 0.4057,
387
+ "step": 315
388
+ },
389
+ {
390
+ "epoch": 2.56,
391
+ "learning_rate": 2.6072059940146775e-06,
392
+ "loss": 0.3894,
393
+ "step": 320
394
+ },
395
+ {
396
+ "epoch": 2.6,
397
+ "learning_rate": 2.1613635589349756e-06,
398
+ "loss": 0.3795,
399
+ "step": 325
400
+ },
401
+ {
402
+ "epoch": 2.64,
403
+ "learning_rate": 1.7555878527937164e-06,
404
+ "loss": 0.3934,
405
+ "step": 330
406
+ },
407
+ {
408
+ "epoch": 2.68,
409
+ "learning_rate": 1.3905907440629752e-06,
410
+ "loss": 0.423,
411
+ "step": 335
412
+ },
413
+ {
414
+ "epoch": 2.72,
415
+ "learning_rate": 1.067012561698319e-06,
416
+ "loss": 0.4121,
417
+ "step": 340
418
+ },
419
+ {
420
+ "epoch": 2.76,
421
+ "learning_rate": 7.854209717842231e-07,
422
+ "loss": 0.4026,
423
+ "step": 345
424
+ },
425
+ {
426
+ "epoch": 2.8,
427
+ "learning_rate": 5.463099816548579e-07,
428
+ "loss": 0.4056,
429
+ "step": 350
430
+ },
431
+ {
432
+ "epoch": 2.84,
433
+ "learning_rate": 3.5009907323737825e-07,
434
+ "loss": 0.3625,
435
+ "step": 355
436
+ },
437
+ {
438
+ "epoch": 2.88,
439
+ "learning_rate": 1.9713246713805588e-07,
440
+ "loss": 0.4279,
441
+ "step": 360
442
+ },
443
+ {
444
+ "epoch": 2.92,
445
+ "learning_rate": 8.767851876239074e-08,
446
+ "loss": 0.3752,
447
+ "step": 365
448
+ },
449
+ {
450
+ "epoch": 2.96,
451
+ "learning_rate": 2.192924752854042e-08,
452
+ "loss": 0.3992,
453
+ "step": 370
454
+ },
455
+ {
456
+ "epoch": 3.0,
457
+ "learning_rate": 0.0,
458
+ "loss": 0.3638,
459
+ "step": 375
460
+ },
461
+ {
462
+ "epoch": 3.0,
463
+ "step": 375,
464
+ "total_flos": 6.763019396186112e+16,
465
+ "train_loss": 0.7424986867904663,
466
+ "train_runtime": 1158.4304,
467
+ "train_samples_per_second": 5.179,
468
+ "train_steps_per_second": 0.324
469
+ }
470
+ ],
471
+ "logging_steps": 5,
472
+ "max_steps": 375,
473
+ "num_train_epochs": 3,
474
+ "save_steps": 100,
475
+ "total_flos": 6.763019396186112e+16,
476
+ "trial_name": null,
477
+ "trial_params": null
478
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75578e2f06f37e3ba4d0e6337a1c276e803dc21ab6505acee9016c03c0e831cf
3
+ size 4792