Wanfq commited on
Commit
31914b9
1 Parent(s): 41c7c5f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -4
README.md CHANGED
@@ -198,7 +198,7 @@ We fine-tune LLaMA-7B with the following hyperparameters:
198
 
199
  | Hyperparameter | Global Batch Size | Learning rate | Epochs | Max length | Weight decay |
200
  |:----------------|-------------------:|---------------:|--------:|------------:|--------------:|
201
- | LLaMA 7B | 128 | 2e-5 | 3 | 2048 | 0 |
202
 
203
  To reproduce the training procedure, please use the following command:
204
 
@@ -259,7 +259,7 @@ python3 ./eval/generate.py \
259
  --prompt_type alpaca \
260
  --do_sample
261
 
262
- 2. Evaluation
263
  python3 ./eval/chatgpt_score.py \
264
  --baseline_file ./eval/answer/<MODEL_1>.jsonl \ # answer of baseline model to compare with
265
  --answer_file ./eval/answer/<MODEL_2>.jsonl \ # answer of evaluation model
@@ -285,10 +285,10 @@ python3 ./eval/generate.py \
285
  --max_new_tokens 512 \
286
  --prompt_type alpaca
287
 
288
- 2. Evaluation
289
  python3 ./eval/auto_eval.py \
290
  --question_file ./eval/question/MATH_eval_set_sample.jsonl \
291
- --answer_file ./eval/answer/<MODEL_ID>.jsonl
292
  ```
293
 
294
  ## Limitations
 
198
 
199
  | Hyperparameter | Global Batch Size | Learning rate | Epochs | Max length | Weight decay |
200
  |:----------------|-------------------:|---------------:|--------:|------------:|--------------:|
201
+ | LLaMA 7B | 128 | 2e-5 | 3 | 512| 0 |
202
 
203
  To reproduce the training procedure, please use the following command:
204
 
 
259
  --prompt_type alpaca \
260
  --do_sample
261
 
262
+ # 2. Evaluation
263
  python3 ./eval/chatgpt_score.py \
264
  --baseline_file ./eval/answer/<MODEL_1>.jsonl \ # answer of baseline model to compare with
265
  --answer_file ./eval/answer/<MODEL_2>.jsonl \ # answer of evaluation model
 
285
  --max_new_tokens 512 \
286
  --prompt_type alpaca
287
 
288
+ # 2. Evaluation
289
  python3 ./eval/auto_eval.py \
290
  --question_file ./eval/question/MATH_eval_set_sample.jsonl \
291
+ --answer_file ./eval/answer/<MODEL_ID>.jsonl # answer of evaluation model
292
  ```
293
 
294
  ## Limitations