cadmonkey-1b-data-2 / config.json
ThomasTheMaker's picture
Upload folder using huggingface_hub
85b3cae verified
{
"model_config": {
"hub_model_name": "ThomasTheMaker/cadmonkey-1b-data-2",
"base_model_name": "unsloth/gemma-3-1b-it-unsloth-bnb-4bit",
"max_seq_length": 1024,
"load_in_4bit": true,
"load_in_8bit": false,
"full_finetuning": false
},
"dataset_config": {
"datasets": [
{
"name": "ThomasTheMaker/Synthetic-Object-v0"
},
{
"name": "ThomasTheMaker/Synthetic-Object-v1"
}
],
"chat_template": "gemma3",
"test_objects": ["cat", "car", "tree"]
},
"lora_config": {
"r": 128,
"alpha_multiplier": 2,
"dropout": 0.05,
"bias": "none",
"use_gradient_checkpointing": "unsloth",
"random_state": 3407,
"use_rslora": false,
"loftq_config": null,
"target_modules": [
"q_proj",
"k_proj",
"v_proj",
"o_proj",
"gate_proj",
"up_proj",
"down_proj"
]
},
"training_config": {
"per_device_eval_batch_size": 0,
"per_device_train_batch_size": 2,
"gradient_accumulation_steps": 8,
"warmup_steps": 5,
"max_steps": -1,
"num_train_epochs": 3,
"learning_rate": 3e-5,
"weight_decay": 0.01,
"lr_scheduler_type": "linear",
"seed": 3407,
"output_dir": "outputs",
"report_to": "none",
"optim": "adamw_8bit",
"logging_steps": 1,
"save_strategy": "steps",
"save_steps": 300,
"save_total_limit": 3,
"eval_strategy": "no",
"eval_steps": 1000
},
"inference_config": {
"max_new_tokens": 1024,
"temperature": 0.8,
"top_p": 0.95,
"top_k": 64,
"do_sample": true
},
"saving_config": {
"save_local": true,
"save_16bit": true,
"save_4bit": false,
"save_lora": true,
"push_to_hub": true
},
"logging_config": {
"csv_log_enabled": true
}
}