"chat_template":"{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
"system_message":"A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.\n\n",
"system_message":"A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions.\n\n",
@ -492,7 +492,7 @@ In this code we provide a flexible way for users to set the conversation templat
On your first run of the data preparation script, you only need to define the "chat_template" (if you want to use custom chat template) and the "system message" (if you want to use a custom system message),
- Step 2: Run the data preparation script--- [prepare_sft_dataset.sh](./examples/data_preparation_scripts/prepare_sft_dataset.sh). Note that whether or not you have skipped the first step, you need to provide the path to the conversation template config file (via the conversation_template_config arg). If you skipped the first step, an auto-generated conversation template will be stored at the designated file path.
- Step 2: Run the data preparation script--- [prepare_sft_dataset.sh](./data_preparation_scripts/prepare_sft_dataset.sh). Note that whether or not you have skipped the first step, you need to provide the path to the conversation template config file (via the conversation_template_config arg). If you skipped the first step, an auto-generated conversation template will be stored at the designated file path.
- Step 3: (Optional) Check the correctness of the processed data. We provided an easy way for you to do a manual checking on the processed data by checking the "$SAVE_DIR/jsonl/part-XXXX.jsonl" files.
@ -512,7 +512,7 @@ Human: <s> what are some pranks with a pen i can do?</s> Assistant: <s> Are you
#### Step 3: Training
Choose a suitable model architecture for your task. Note that your model should be compatible with the tokenizer that you used to tokenize the SFT dataset. You can run [train_sft.sh](./examples/training_scripts/train_sft.sh) to start a supervised instructs fine-tuning. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
Choose a suitable model architecture for your task. Note that your model should be compatible with the tokenizer that you used to tokenize the SFT dataset. You can run [train_sft.sh](./training_scripts/train_sft.sh) to start a supervised instructs fine-tuning. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
### RLHF Training Stage2 - Training Reward Model
@ -554,11 +554,11 @@ Below shows the preference dataset format used in training the reward model.
#### Step 2: Preprocessing
Similar to the second step in the previous stage, we format the reward data into the same structured format as used in step 2 of the SFT stage. You can run [prepare_preference_dataset.sh](./examples/data_preparation_scripts/prepare_preference_dataset.sh) to prepare the preference data for reward model training.
Similar to the second step in the previous stage, we format the reward data into the same structured format as used in step 2 of the SFT stage. You can run [prepare_preference_dataset.sh](./data_preparation_scripts/prepare_preference_dataset.sh) to prepare the preference data for reward model training.
#### Step 3: Training
You can run [train_rm.sh](./examples/training_scripts/train_rm.sh) to start the reward model training. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
You can run [train_rm.sh](./training_scripts/train_rm.sh) to start the reward model training. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
#### Features and Tricks in RM Training
@ -629,14 +629,14 @@ The second dataset--- pretrained dataset is optional, provide it if you want to
]
```
#### Step 2: Preprocessing
To prepare the prompt dataset for PPO training, simply run [prepare_prompt_dataset.sh](./examples/data_preparation_scripts/prepare_prompt_dataset.sh)
To prepare the prompt dataset for PPO training, simply run [prepare_prompt_dataset.sh](./data_preparation_scripts/prepare_prompt_dataset.sh)
You can use the SFT dataset you prepared in the SFT stage or prepare a new one from different source for the ptx dataset. The ptx data is used to calculate ptx loss, which stabilizes the training according to the [InstructGPT paper](https://arxiv.org/pdf/2203.02155.pdf).
#### Step 3: Training
You can run the [train_ppo.sh](./examples/training_scripts/train_ppo.sh) to start PPO training. Here are some unique arguments for PPO, please refer to the training configuration section for other training configuration. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
You can run the [train_ppo.sh](./training_scripts/train_ppo.sh) to start PPO training. Here are some unique arguments for PPO, please refer to the training configuration section for other training configuration. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options.
```bash
@ -720,7 +720,7 @@ For DPO training, you only need the preference dataset. Please follow the instru
#### Step 2: Training
You can run the [train_dpo.sh](./examples/training_scripts/train_dpo.sh) to start DPO training. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options. Following the trend of recent research on DPO-like alignment methods, we added option for the user to choose from, including whether to do length normalization , reward shaping and whether to use a reference model in calculating implicit reward. Here are those options,
You can run the [train_dpo.sh](./training_scripts/train_dpo.sh) to start DPO training. Please refer to the [training configuration](#training-configuration) section for details regarding supported training options. Following the trend of recent research on DPO-like alignment methods, we added option for the user to choose from, including whether to do length normalization , reward shaping and whether to use a reference model in calculating implicit reward. Here are those options,
```
--beta 0.1 \ # the temperature in DPO loss, Default to 0.1
@ -737,7 +737,7 @@ You can run the [train_dpo.sh](./examples/training_scripts/train_dpo.sh) to star
### Alternative Option For RLHF: Simple Preference Optimization
We support the method introduced in the paper [SimPO: Simple Preference Optimization
with a Reference-Free Reward](https://arxiv.org/pdf/2405.14734) (SimPO). Which is a reference model free aligment method that add length normalization and reward shaping to the DPO loss to enhance training stability and efficiency. As the method doesn't deviate too much from DPO, we add support for length normalization and SimPO reward shaping in our DPO implementation. To use SimPO in alignment, use the [train_dpo.sh](./examples/training_scripts/train_dpo.sh) script, set the `loss_type` to `simpo_loss`, you can also set the value for temperature (`beta`) and reward target margin (`gamma`) but it is optional.
with a Reference-Free Reward](https://arxiv.org/pdf/2405.14734) (SimPO). Which is a reference model free aligment method that add length normalization and reward shaping to the DPO loss to enhance training stability and efficiency. As the method doesn't deviate too much from DPO, we add support for length normalization and SimPO reward shaping in our DPO implementation. To use SimPO in alignment, use the [train_dpo.sh](./training_scripts/train_dpo.sh) script, set the `loss_type` to `simpo_loss`, you can also set the value for temperature (`beta`) and reward target margin (`gamma`) but it is optional.
#### SimPO Result
<palign="center">
@ -746,7 +746,7 @@ with a Reference-Free Reward](https://arxiv.org/pdf/2405.14734) (SimPO). Which i
### Alternative Option For RLHF: Odds Ratio Preference Optimization
We support the method introduced in the paper [ORPO: Monolithic Preference Optimization without Reference Model](https://arxiv.org/abs/2403.07691) (ORPO). Which is a reference model free aligment method that mixes the SFT loss with a reinforcement learning loss that uses odds ratio as the implicit reward to enhance training stability and efficiency. To use ORPO in alignment, use the [train_orpo.sh](./examples/training_scripts/train_orpo.sh) script, You can set the value for `lambda` (which determine how strongly the reinforcement learning loss affect the training) but it is optional.
We support the method introduced in the paper [ORPO: Monolithic Preference Optimization without Reference Model](https://arxiv.org/abs/2403.07691) (ORPO). Which is a reference model free aligment method that mixes the SFT loss with a reinforcement learning loss that uses odds ratio as the implicit reward to enhance training stability and efficiency. To use ORPO in alignment, use the [train_orpo.sh](./training_scripts/train_orpo.sh) script, You can set the value for `lambda` (which determine how strongly the reinforcement learning loss affect the training) but it is optional.
Tong Li
4 months ago
committed by