mirror of
https://github.com/wassname/Open-Assistant.git
synced 2026-06-28 16:20:34 +08:00
theblackcat102
74 lines
1.6 KiB
Markdown
74 lines
1.6 KiB
Markdown
# Train using supervised examples
|
|
|
|
Requirements
|
|
|
|
```
|
|
wandb
|
|
evaluate
|
|
datasets
|
|
transformers
|
|
torch
|
|
```
|
|
|
|
Start training reward model
|
|
|
|
```bash
|
|
python trainer.py --configs defaults galactica-125
|
|
```
|
|
|
|
## Dataset
|
|
|
|
For now we only support webgpt and summary dataset from OpenAI. Once
|
|
open-asisstant dataset are available it will be added here.
|
|
|
|
## Model
|
|
|
|
Normally you should be able to add new models in configs/config.yml
|
|
|
|
```
|
|
your-model-name:
|
|
learning_rate: 2e-6
|
|
model_name: <huggingface model name>
|
|
weight_decay: 0.01
|
|
max_length: 812
|
|
warmup_steps: 600
|
|
gradient_checkpointing: false
|
|
gradient_accumulation_steps: 5
|
|
per_device_train_batch_size: 4
|
|
per_device_eval_batch_size: 4
|
|
```
|
|
|
|
```
|
|
python trainer.py --configs defaults your-model-name
|
|
```
|
|
|
|
However, if the model of your choice doesn't have pad_token, eos_token,
|
|
sep_token, you have to update utils.py `get_tokenizer` to use the right token.
|
|
|
|
## Deepspeed support
|
|
|
|
You can edit the configs/zero_config.json and use any stage you wish. The
|
|
current config uses zero-stage 3. For more details on how to setup the config
|
|
checkout [this page](https://www.deepspeed.ai/tutorials/zero/)
|
|
|
|
Once you are satisfy with your deepzero config, you can add --deepspeed flag at
|
|
the end to trigger deepspeed
|
|
|
|
```
|
|
python trainer.py --configs defaults your-model-name --deepspeed
|
|
```
|
|
|
|
## Dataset choices
|
|
|
|
## Results
|
|
|
|
Experimental results in wandb
|
|
[here](https://wandb.ai/sanagnos/supervised-finetuning?workspace=user-sanagnos).
|
|
|
|
## TODOS
|
|
|
|
- decide on a model
|
|
- Merge utils etc with reward model
|
|
- Casual Modelling for GPT-JT does not leverage the bidirectional mask for the
|
|
prompt? (https://huggingface.co/togethercomputer/GPT-JT-6B-v1)
|