PEFT-method-comparison / text_generation_benchmark /default_benchmark_params.json
github-actions[bot]
🚀 Deploy method comparison app from GH action
08d2048
raw
history blame contribute delete
308 Bytes
{
"model_id": "meta-llama/Llama-3.2-3B",
"dtype": "float16",
"seed": 42,
"num_inference_runs": 10,
"max_new_tokens": 20,
"category_generation_params": {
"short": {"max_new_tokens": 20},
"medium": {"max_new_tokens": 50},
"long": {"max_new_tokens": 100}
}
}