neuralmagic
/

Meta-Llama-3.1-70B-Instruct-quantized.w8a16

Text Generation

text-generation-inference

Inference Endpoints

compressed-tensors

Model card Files Files and versions Community

Meta-Llama-3.1-70B-Instruct-quantized.w8a16

File size: 173 Bytes

3d7e3dc

quant_stage:
  quant_modifiers:
    GPTQModifier:
      sequential_update: false
      dampening_frac: 0.1
      ignore: [lm_head]
      scheme: W8A16
      targets: Linear