| ## Mixtral-8x7B-Instruct-v0.1-FP8-v2 | |
| * Weights and activations are per-tensor quantized to float8_e4m3. | |
| * Quantization with AutoFP8. | |
| * Calibration dataset: Ultrachat-200k | |
| * Samples: 4096 | |
| * Sequence length: 8192 | |
| ## Evaluation | |
| TBA |
| ## Mixtral-8x7B-Instruct-v0.1-FP8-v2 | |
| * Weights and activations are per-tensor quantized to float8_e4m3. | |
| * Quantization with AutoFP8. | |
| * Calibration dataset: Ultrachat-200k | |
| * Samples: 4096 | |
| * Sequence length: 8192 | |
| ## Evaluation | |
| TBA |