| ## Mixtral-8x22B-Instruct-v0.1-FP8-v1 | |
| * Weights and activations are per-tensor quantized to float8_e4m3. | |
| * Quantization with AutoFP8. | |
| * Calibration dataset: Ultrachat (mgoin/ultrachat_2k) | |
| * Samples: 2048 | |
| * Sequence length: 8192 | |
| ## Evaluation | |
| TBA |
| ## Mixtral-8x22B-Instruct-v0.1-FP8-v1 | |
| * Weights and activations are per-tensor quantized to float8_e4m3. | |
| * Quantization with AutoFP8. | |
| * Calibration dataset: Ultrachat (mgoin/ultrachat_2k) | |
| * Samples: 2048 | |
| * Sequence length: 8192 | |
| ## Evaluation | |
| TBA |