mtasic85 commited on
Commit
1ff18cb
·
1 Parent(s): 7224ded

cpt core 4

Browse files
Files changed (2) hide show
  1. README.md +1 -1
  2. scripts/cpt_core_model_4.py +2 -3
README.md CHANGED
@@ -407,7 +407,7 @@ python -B convert_pth_to_safetensors.py
407
  ```
408
 
409
  ```bash
410
- python -B cpt_core_model_4.py
411
  ```
412
 
413
  ```
 
407
  ```
408
 
409
  ```bash
410
+ CUDA_VISIBLE_DEVICES=0 python -B cpt_core_model_4.py
411
  ```
412
 
413
  ```
scripts/cpt_core_model_4.py CHANGED
@@ -80,12 +80,11 @@ def unlsoth_generator():
80
  global litgpt_streaming_dataset
81
 
82
  for batch in litgpt_streaming_dataset:
83
- # print(batch)
84
  yield {'input_ids': batch}
85
- break
86
 
87
 
88
- train_dataset = Dataset.from_generator(unlsoth_generator, streaming=True)
 
89
 
90
 
91
  from trl import SFTTrainer
 
80
  global litgpt_streaming_dataset
81
 
82
  for batch in litgpt_streaming_dataset:
 
83
  yield {'input_ids': batch}
 
84
 
85
 
86
+ # train_dataset = Dataset.from_generator(unlsoth_generator, streaming=True)
87
+ train_dataset = Dataset.from_generator(unlsoth_generator, batched=True)
88
 
89
 
90
  from trl import SFTTrainer