|
{ |
|
"comp_cgenerate_active": false, |
|
"comp_ctranslate_active": false, |
|
"comp_cwhisper_active": false, |
|
"comp_diffusers2_active": false, |
|
"comp_ifw_active": false, |
|
"comp_onediff_active": false, |
|
"comp_step_caching_active": false, |
|
"comp_torch_compile_active": false, |
|
"comp_ws2t_active": false, |
|
"comp_x-fast_active": false, |
|
"prune_torch-structured_active": false, |
|
"quant_aqlm_active": false, |
|
"quant_awq_active": false, |
|
"quant_gptq_active": false, |
|
"quant_half_active": false, |
|
"quant_hqq_active": false, |
|
"quant_llm-int8_active": true, |
|
"quant_quanto_active": false, |
|
"quant_torch_dynamic_active": false, |
|
"quant_torch_static_active": false, |
|
"quant_llm-int8_compute_dtype": "bfloat16", |
|
"quant_llm-int8_double_quant": false, |
|
"quant_llm-int8_enable_fp32_cpu_offload": false, |
|
"quant_llm-int8_has_fp16_weight": false, |
|
"quant_llm-int8_quant_type": "fp4", |
|
"quant_llm-int8_threshold": 6.0, |
|
"quant_llm-int8_weight_bits": 8, |
|
"max_batch_size": 1, |
|
"device": "cuda", |
|
"cache_dir": "/covalent/.cache/models/tmp2rmhcffl", |
|
"task": "", |
|
"save_load_fn": "bitsandbytes", |
|
"save_load_fn_args": { |
|
"weight_quantization_bits": "param.dtype" |
|
}, |
|
"api_key": null |
|
} |