ComfyUI Node: T5 Quantization Config 🐼
Authored by zhongpei
Created 9 months ago
Updated 5 months ago
272 stars
Category
fofo🐼/prompt
Inputs
quantization_mode
- none
- load_in_8bit
- load_in_4bit
llm_int8_threshold FLOAT
llm_int8_skip_modules STRING
llm_int8_enable_fp32_cpu_offload BOOLEAN
llm_int8_has_fp16_weight BOOLEAN
bnb_4bit_compute_dtype STRING
bnb_4bit_quant_type STRING
bnb_4bit_use_double_quant BOOLEAN
bnb_4bit_quant_storage STRING
Outputs
QUANTIZATIONCONFIG
Extension: Comfyui_image2prompt
Nodes:Image to Text, Loader Image to Text Model.
Authored by zhongpei