Mahou-1.0-llama3-8B / README.md
nbeerbower's picture
Update README.md
1cc918d verified
|
raw
history blame
2.24 kB
---
library_name: transformers
tags: []
base_model:
- nbeerbower/llama-3-stella-truthy-dpo-8B
datasets:
- flammenai/FlameMix-DPO-v1
license: other
license_name: llama3
---
![image/png](https://huggingface.co/flammenai/Mahou-1.0-mistral-7B/resolve/main/mahou1.png)
# Mahou-1.0-llama3
Mahou is our attempt to build a production-ready conversational/roleplay LLM.
Future versions will be released iteratively and finetuned from flammen.ai conversational data.
NOTE: this model is experimental and currently significantly flawed.
### License
This model is based on Meta Llama-3-8B and is governed by the [META LLAMA 3 COMMUNITY LICENSE AGREEMENT](LICENSE).
### Method
Finetuned using an A100 on Google Colab.
[Fine-tune a Mistral-7b model with Direct Preference Optimization](https://towardsdatascience.com/fine-tune-a-mistral-7b-model-with-direct-preference-optimization-708042745aac) - [Maxime Labonne](https://huggingface.co/mlabonne)
### Configuration
LoRA, model, and training settings:
```python
# LoRA configuration
peft_config = LoraConfig(
r=16,
lora_alpha=16,
lora_dropout=0.05,
bias="none",
task_type="CAUSAL_LM",
target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']
)
# Model to fine-tune
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.bfloat16,
load_in_4bit=True
)
model.config.use_cache = False
# Reference model
ref_model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.bfloat16,
load_in_4bit=True
)
# Training arguments
training_args = TrainingArguments(
per_device_train_batch_size=4,
gradient_accumulation_steps=4,
gradient_checkpointing=True,
learning_rate=5e-5,
lr_scheduler_type="cosine",
max_steps=2000,
save_strategy="no",
logging_steps=1,
output_dir=new_model,
optim="paged_adamw_32bit",
warmup_steps=100,
bf16=True,
report_to="wandb",
)
# Create DPO trainer
dpo_trainer = DPOTrainer(
model,
ref_model,
args=training_args,
train_dataset=dataset,
tokenizer=tokenizer,
peft_config=peft_config,
beta=0.1,
max_prompt_length=2048,
max_length=8192,
force_use_ref_model=True
)
```