| { | |
| "base_model": "unsloth/llama-3-8b-bnb-4bit", | |
| "dataset": "metythorn/khmer-corpus", | |
| "training_framework": "unsloth", | |
| "lora_config": { | |
| "r": 128, | |
| "lora_alpha": 32, | |
| "lora_dropout": 0, | |
| "bias": "none", | |
| "target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "gate_proj", | |
| "up_proj", | |
| "down_proj", | |
| "embed_proj", | |
| "lm_head" | |
| ], | |
| "use_rslora": true | |
| }, | |
| "training_args": { | |
| "per_device_train_batch_size": 2, | |
| "gradient_accumulation_steps": 8, | |
| "warmup_ratio": 0.1, | |
| "num_train_epochs": 1, | |
| "learning_rate": 5e-05, | |
| "embedding_learning_rate": 5e-06, | |
| "optim": "adamw_8bit", | |
| "lr_scheduler_type": "cosine", | |
| "weight_decay": 0.0, | |
| "seed": 3407 | |
| }, | |
| "model_config": { | |
| "max_seq_length": 2048, | |
| "load_in_4bit": true, | |
| "use_gradient_checkpointing": "unsloth" | |
| } | |
| } |