| default_stage: | |
| default_modifiers: | |
| QuantizationModifier: | |
| targets: [Linear] | |
| ignore: [model.embed_tokens, embed_tokens, lm_head, 're:.*layernorm.*', 're:.*layer_norm.*', | |
| 're:.*norm.*', 're:.*rotary.*'] | |
| scheme: FP8_DYNAMIC | |
| default_stage: | |
| default_modifiers: | |
| QuantizationModifier: | |
| targets: [Linear] | |
| ignore: [model.embed_tokens, embed_tokens, lm_head, 're:.*layernorm.*', 're:.*layer_norm.*', | |
| 're:.*norm.*', 're:.*rotary.*'] | |
| scheme: FP8_DYNAMIC | |