amd / Mixtral-8x22B-Instruct-v0.1-FP8-KV

huggingface.co
Total runs: 1.9K
24-hour runs: -76
7-day runs: -142
30-day runs: -989
Model's Last Updated: December 20 2024

Introduction of Mixtral-8x22B-Instruct-v0.1-FP8-KV

Model Details of Mixtral-8x22B-Instruct-v0.1-FP8-KV

Mixtral-8x22B-Instruct-v0.1-FP8-KV

  • Introduction
    This model was created by applying Quark with calibration samples from Pile dataset.
  • Quantization Stragegy
    • Quantized Layers : All linear layers excluding "lm_head", "*.gate"
    • Weight : FP8 symmetric per-tensor
    • Activation : FP8 symmetric per-tensor
    • KV Cache : FP8 symmetric per-tensor
  • Quick Start
  1. Download and install Quark
  2. Run the quantization script in the example folder using the following command line:
export MODEL_DIR = [local model checkpoint folder] or mistralai/Mixtral-8x22B-Instruct-v0.1
# single GPU
python3 quantize_quark.py \
        --model_dir $MODEL_DIR \
        --output_dir Mixtral-8x22B-Instruct-v0.1-FP8-KV \
        --quant_scheme w_fp8_a_fp8 \
        --kv_cache_dtype fp8 \
        --num_calib_data 128 \
        --model_export quark_safetensors \
        --no_weight_matrix_merge
# If model size is too large for single GPU, please use multi GPU instead.
python3 quantize_quark.py \
        --model_dir $MODEL_DIR \
        --output_dir Mixtral-8x22B-Instruct-v0.1-FP8-KV \
        --quant_scheme w_fp8_a_fp8 \
        --kv_cache_dtype fp8 \
        --num_calib_data 128 \
        --model_export quark_safetensors \
        --no_weight_matrix_merge \
        --multi_gpu
Deployment

Quark has its own export format and allows FP8 quantized models to be efficiently deployed using the vLLM backend(vLLM-compatible).

Evaluation

Quark currently uses perplexity(PPL) as the evaluation metric for accuracy loss before and after quantization.The specific PPL algorithm can be referenced in the quantize_quark.py. The quantization evaluation results are conducted in pseudo-quantization mode, which may slightly differ from the actual quantized inference accuracy. These results are provided for reference only.

Evaluation scores
Benchmark Mixtral-8x22B-Instruct-v0.1 Mixtral-8x22B-Instruct-v0.1-FP8-KV(this model)
Perplexity-wikitext2 2.8871 2.9193
License

Modifications copyright(c) 2024 Advanced Micro Devices,Inc. All rights reserved.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Runs of amd Mixtral-8x22B-Instruct-v0.1-FP8-KV on huggingface.co

1.9K
Total runs
-76
24-hour runs
-71
3-day runs
-142
7-day runs
-989
30-day runs

More Information About Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co Model

More Mixtral-8x22B-Instruct-v0.1-FP8-KV license Visit here:

https://choosealicense.com/licenses/apache-2.0

Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co

Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co is an AI model on huggingface.co that provides Mixtral-8x22B-Instruct-v0.1-FP8-KV's model effect (), which can be used instantly with this amd Mixtral-8x22B-Instruct-v0.1-FP8-KV model. huggingface.co supports a free trial of the Mixtral-8x22B-Instruct-v0.1-FP8-KV model, and also provides paid use of the Mixtral-8x22B-Instruct-v0.1-FP8-KV. Support call Mixtral-8x22B-Instruct-v0.1-FP8-KV model through api, including Node.js, Python, http.

Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co Url

https://huggingface.co/amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV

amd Mixtral-8x22B-Instruct-v0.1-FP8-KV online free

Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co is an online trial and call api platform, which integrates Mixtral-8x22B-Instruct-v0.1-FP8-KV's modeling effects, including api services, and provides a free online trial of Mixtral-8x22B-Instruct-v0.1-FP8-KV, you can try Mixtral-8x22B-Instruct-v0.1-FP8-KV online for free by clicking the link below.

amd Mixtral-8x22B-Instruct-v0.1-FP8-KV online free url in huggingface.co:

https://huggingface.co/amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV

Mixtral-8x22B-Instruct-v0.1-FP8-KV install

Mixtral-8x22B-Instruct-v0.1-FP8-KV is an open source model from GitHub that offers a free installation service, and any user can find Mixtral-8x22B-Instruct-v0.1-FP8-KV on GitHub to install. At the same time, huggingface.co provides the effect of Mixtral-8x22B-Instruct-v0.1-FP8-KV install, users can directly use Mixtral-8x22B-Instruct-v0.1-FP8-KV installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

Mixtral-8x22B-Instruct-v0.1-FP8-KV install url in huggingface.co:

https://huggingface.co/amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV

Url of Mixtral-8x22B-Instruct-v0.1-FP8-KV

Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co Url

Provider of Mixtral-8x22B-Instruct-v0.1-FP8-KV huggingface.co

amd
ORGANIZATIONS

Other API from amd

huggingface.co

Total runs: 184.5K
Run Growth: 97.3K
Growth Rate: 52.72%
Updated:April 14 2026
huggingface.co

Total runs: 114.2K
Run Growth: -5.4K
Growth Rate: -4.76%
Updated:July 17 2026
huggingface.co

Total runs: 104.0K
Run Growth: 12.6K
Growth Rate: 12.15%
Updated:June 19 2026
huggingface.co

Total runs: 52.6K
Run Growth: -5.8K
Growth Rate: -11.01%
Updated:July 01 2026
huggingface.co

Total runs: 24.6K
Run Growth: 13.9K
Growth Rate: 56.56%
Updated:July 27 2026
huggingface.co

Total runs: 12.4K
Run Growth: 4.6K
Growth Rate: 37.39%
Updated:October 09 2024
huggingface.co

Total runs: 11.1K
Run Growth: 1.2K
Growth Rate: 10.43%
Updated:August 12 2025
huggingface.co

Total runs: 2.6K
Run Growth: -15
Growth Rate: -0.59%
Updated:June 19 2026
huggingface.co

Total runs: 1.3K
Run Growth: -37.3K
Growth Rate: -2875.40%
Updated:June 19 2026
huggingface.co

Total runs: 987
Run Growth: -2.8K
Growth Rate: -295.15%
Updated:May 27 2026
huggingface.co

Total runs: 547
Run Growth: -4.4K
Growth Rate: -812.25%
Updated:June 19 2026
huggingface.co

Total runs: 512
Run Growth: -396
Growth Rate: -77.34%
Updated:November 15 2025