Instructions to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16 # Run inference directly in the terminal: llama cli -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16 # Run inference directly in the terminal: llama cli -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16 # Run inference directly in the terminal: ./build/bin/llama cli -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Use Docker
docker model run hf.co/webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
- Llama new
- LM Studio
- Jan
- vLLM
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
- Ollama
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with Ollama:
ollama run hf.co/webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
- Unsloth Desktop
- Pi
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:9931/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with Docker Model Runner:
docker model run hf.co/webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
- Lemonade
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Run and chat with the model
lemonade run user.Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF-F16
List all available models
lemonade list
- Hermes Agent
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:9931/v1 hermes config set model.default webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:9931/v1 \ --custom-model-id "webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF:F16" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Sakura โ MiMo-V2.6-Distill-Qwen-9B (GGUF, three sizes)
MiMo-V2.6-Distill-Qwen-9B is a 9B agentic model from Xiaomi MiMo: Qwen3.5-9B fine-tuned on MiMo-generated data for coding, tool use and general agent tasks (multimodal Qwen3.5 architecture). This repository holds three GGUF files of different size made from the BF16 weights with our measured mixed-codec method. Pick one on the Files page; the table below says what each file is. Independent community quantization, not an official Xiaomi MiMo release. Part of the Sakura Mini line.
Quality at a glance
Same method, same texts, same machine for every file; lower KL divergence means closer to the BF16 original:
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.ggufvs bartowski IQ3_XXS (3.85 GiB): KL divergence to the BF16 original 7 % lower on average over all 3 texts (0.07 GiB smaller).Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.ggufvs bartowski IQ4_XS (4.87 GiB): KL divergence to the BF16 original 3 % lower on average over all 3 texts (0.05 GiB smaller).Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.ggufvs bartowski Q4_K_M (5.44 GiB): KL divergence to the BF16 original 42 % lower on average over all 3 texts (0.11 GiB larger).
Why: the bit budget per tensor is allocated from measured sensitivity (see How it was made), not from fixed rules. Full table below.
Which file should I take?
| File | For |
|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf (3.78 GiB, 3.62 bpw) |
smallest: for tight memory budgets; expect visible quality loss |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf (4.81 GiB, 4.61 bpw) |
middle: best balance of size and quality for most machines |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf (5.54 GiB, 5.31 bpw) |
largest: closest to the BF16 original of the three |
The files
| File | Size | bits/weight | KLD en | KLD dev | KLD wiki | same top token (avg) | Tensor types by size |
|---|---|---|---|---|---|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf |
3.78 GiB | 3.62 | 0.0642 | 0.0628 | 0.1146 | 89.2 % | IQ3_S 55%, IQ4_XS 25%, Q4_K 9%, IQ2_S 8%, Q5_K 3% |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf |
4.81 GiB | 4.61 | 0.0179 | 0.0194 | 0.0285 | 94.3 % | Q4_K 58%, IQ4_XS 22%, Q5_K 20% |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf |
5.54 GiB | 5.31 | 0.0074 | 0.0076 | 0.0103 | 96.4 % | Q5_K 83%, IQ4_XS 9%, Q4_K 6% |
"Tensor types by size" lists the share of the file's bytes per ggml type (all are standard llama.cpp types, so the files run in stock llama.cpp). Quality columns: mean KL divergence of the quantized model's next-token distribution against the BF16 source on held-out text (lower is better) and how often the most likely token is unchanged (higher is better).
mmproj-MiMo-V2.6-Distill-Qwen-9B-f16.gguf: vision projector (unchanged from bartowski's F16 conversion).
Comparison at similar size
Same texts, same context, same chunks, measured by us with llama-perplexity (BF16 source = reference; PPL of the BF16 model: en: 1.988, dev: 2.187, wiki: 8.688). The reference quants are: bartowski (imatrix), measured the same way. Sorted by size.
| Quant | Source | Size | KLD en | KLD dev | KLD wiki | PPL en | same top token (avg) |
|---|---|---|---|---|---|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf |
this repository | 3.79 GiB | 0.0642 | 0.0628 | 0.1146 | 2.065 | 89.2 % |
| IQ3_XXS | bartowski (imatrix) | 3.85 GiB | 0.0714 | 0.0662 | 0.1215 | 2.027 | 89.7 % |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf |
this repository | 4.82 GiB | 0.0179 | 0.0194 | 0.0285 | 2.020 | 94.3 % |
| IQ4_XS | bartowski (imatrix) | 4.87 GiB | 0.0198 | 0.0188 | 0.0290 | 2.013 | 94.6 % |
| Q4_K_M | bartowski (imatrix) | 5.44 GiB | 0.0127 | 0.0119 | 0.0200 | 2.007 | 95.5 % |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf |
this repository | 5.55 GiB | 0.0074 | 0.0076 | 0.0103 | 2.014 | 96.4 % |
| Q5_K_M | bartowski (imatrix) | 6.40 GiB | 0.0039 | 0.0034 | 0.0059 | 2.011 | 97.4 % |
Method: 12 chunks of 512 tokens per text; texts: English held-out text, developer-text held-out set, English encyclopedia held-out text (not used in any calibration). Single measurement on one machine; small KLD differences at similar size are not a quality ranking.
How it was made
- BF16 GGUF of the original model and the importance matrix come from bartowski/MiMo-V2.6-Distill-Qwen-9B-GGUF; we used both unchanged (credit and thanks to bartowski).
- Every large weight matrix was quantized once per candidate type (Q2_K, IQ2_S, IQ3_XXS, IQ3_S, Q3_K, IQ4_XS, Q4_K, Q5_K, Q6_K, Q8_0) with
llama-quantizeand the importance matrix; the error of each choice was estimated per matrix (importance-weighted) and scaled by a sensitivity factor measured with real KL divergence: every group of tensors (for example the FFN down projections, the attention value projections, the first and last layers) was moved alone to a lower-bit type, and the KL divergence it caused was compared with its quantization error. - An exact budget allocation (multiple-choice knapsack over bytes) picked one type per matrix for each target size; the final model was assembled from the stored tensors without re-quantizing. We do not publish per-tensor choices. Norms, small tensors and the multi-token-prediction layers stay at high precision.
Limits
- Measured only with KL divergence and perplexity on short held-out texts and not with downstream benchmarks; do not read it as a task-quality claim.
- Quantization always costs some quality; the smallest file costs the most.
- The vision projector
mmproj-MiMo-V2.6-Distill-Qwen-9B-f16.ggufis bartowski's conversion (unchanged, F16); it is needed for images and is not quantized by us.
Credits
- Original model: MiMo-V2.6-Distill-Qwen-9B (Xiaomi MiMo, MIT), MIT license (included as
LICENSE). - BF16 GGUF, importance matrix and the reference quants: bartowski.
- GGUF format and tools: ggml-org/llama.cpp (MIT).
ไธญๆ่ฏดๆ ยท ๆจฑ่ฑ (Simplified Chinese)
English above. ๆฌ่ไธบไธๆ็ไธญๆ็ฟป่ฏ(Sakura = ๆจฑ่ฑ yฤซnghuฤ);ๅฎๆด็็ฌ็ซไธญๆ็่ง README_zh.mdใ
Sakura โ MiMo-V2.6-Distill-Qwen-9B(GGUF,ไธ็งๅคงๅฐ)
MiMo-V2.6-Distill-Qwen-9B ๆฏ Xiaomi MiMo ็ไธไธช 9B ๆบ่ฝไฝๆจกๅ:Qwen3.5-9B ๅจ MiMo ็ๆ็ๆฐๆฎไธ้ๅฏน็ผ็จใๅทฅๅ ทไฝฟ็จๅ้็จๆบ่ฝไฝไปปๅก่ฟ่กไบๅพฎ่ฐ(ๅคๆจกๆ Qwen3.5 ๆถๆ)ใๆฌไปๅบๅ ๅซ ไธไธชไธๅๅคงๅฐ็ GGUF ๆไปถ,็ฑ BF16 ๆ้้่ฟๆไปฌๅฎๆตๅพๅบ็ๆททๅ็ผ็ ๆนๆณๅถไฝใ่ฏทๅจ Files ้กต้ข้ๆฉๅ ถไธญไธไธช;ไธ่กจ่ฏดๆไบๆฏไธชๆไปถ็็จ้ใ ่ฟๆฏ็ฌ็ซ็็คพๅบ้ๅ็ๆฌ,ไธๆฏ Xiaomi MiMo ็ๅฎๆนๅๅธใๅฑไบ Sakura Mini ็ณปๅใ
่ดจ้ๆฆ่ง
ๆฏไธชๆไปถไฝฟ็จ็ธๅ็ๆนๆณใ็ธๅ็ๆๆฌใๅไธๅฐๆบๅจ;KL ๆฃๅบฆ่ถไฝ,่ถๆฅ่ฟ BF16 ๅๅงๆจกๅ:
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.ggufๅฏนๆฏ bartowski IQ3_XXS (3.85 GiB):็ธๅฏนไบ BF16 ๅๅงๆจกๅ,ๅจๅ จ้จ 3 ไธชๆๆฌไธๅนณๅ KL ๆฃๅบฆ ไฝ 7 %(ๅฐ 0.07 GiB)ใSakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.ggufๅฏนๆฏ bartowski IQ4_XS (4.87 GiB):็ธๅฏนไบ BF16 ๅๅงๆจกๅ,ๅจๅ จ้จ 3 ไธชๆๆฌไธๅนณๅ KL ๆฃๅบฆ ไฝ 3 %(ๅฐ 0.05 GiB)ใSakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.ggufๅฏนๆฏ bartowski Q4_K_M (5.44 GiB):็ธๅฏนไบ BF16 ๅๅงๆจกๅ,ๅจๅ จ้จ 3 ไธชๆๆฌไธๅนณๅ KL ๆฃๅบฆ ไฝ 42 %(ๅคง 0.11 GiB)ใ
ๅๅ :ๆฏไธชๅผ ้็ๆฏ็น้ข็ฎๆฏๆ นๆฎๅฎๆต็ๆๆๅบฆๅ้ ็(่ง ๅถไฝๆนๅผ),่ไธๆฏๆๅบๅฎ่งๅใๅฎๆด่กจๆ ผ่งไธๆใ
ๆ่ฏฅ้ๅชไธชๆไปถ?
| ๆไปถ | ้็จๅบๆฏ |
|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf(3.78 GiB,3.62 bpw) |
ๆๅฐ:้ๅๅ ๅญ้ข็ฎ็ดงๅผ ็ๆ ๅต;้ข่ฎก่ดจ้ไผๆๆๆพๆๅคฑ |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf(4.81 GiB,4.61 bpw) |
ๅฑ ไธญ:ๅฏนๅคงๅคๆฐๆบๅจ่่จ,ๅคงๅฐไธ่ดจ้็ๆไฝณๅนณ่กก |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf(5.54 GiB,5.31 bpw) |
ๆๅคง:ไธ่ ไธญๆๆฅ่ฟ BF16 ๅๅงๆจกๅ |
่ฟไบๆไปถ
| ๆไปถ | ๅคงๅฐ | bits/weight | KLD en | KLD dev | KLD wiki | same top token (avg) | ๅๅผ ้็ฑปๅ(ๆๅคงๅฐ) |
|---|---|---|---|---|---|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf |
3.78 GiB | 3.62 | 0.0642 | 0.0628 | 0.1146 | 89.2 % | IQ3_S 55%, IQ4_XS 25%, Q4_K 9%, IQ2_S 8%, Q5_K 3% |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf |
4.81 GiB | 4.61 | 0.0179 | 0.0194 | 0.0285 | 94.3 % | Q4_K 58%, IQ4_XS 22%, Q5_K 20% |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf |
5.54 GiB | 5.31 | 0.0074 | 0.0076 | 0.0103 | 96.4 % | Q5_K 83%, IQ4_XS 9%, Q4_K 6% |
โๅๅผ ้็ฑปๅ(ๆๅคงๅฐ)โๅๅบๆฏไธช ggml ็ฑปๅๅ ๆไปถๅญ่ๆฐ็ๆฏไพ(ๅ จ้จๆฏๆ ๅ็ llama.cpp ็ฑปๅ,ๅ ๆญค่ฟไบๆไปถๅฏๅจๅ็ llama.cpp ไธญ่ฟ่ก)ใ่ดจ้ๅ:้ๅๆจกๅ็ไธไธไธช token ๅๅธ็ธๅฏนไบ BF16 ๆบๆไปถๅจ็ๅบๆๆฌไธ็ๅนณๅ KL ๆฃๅบฆ(่ถไฝ่ถๅฅฝ),ไปฅๅๆๅฏ่ฝ็ token ไฟๆไธๅ็้ข็(่ถ้ซ่ถๅฅฝ)ใ
mmproj-MiMo-V2.6-Distill-Qwen-9B-f16.gguf:่ง่งๆๅฝฑๅจ(ไธ bartowski ็ F16 ่ฝฌๆข็ๆฌ็ธๅ,ๆชๆนๅจ)ใ
็ธ่ฟๅคงๅฐ็ๆฏ่พ
็ธๅ็ๆๆฌใ็ธๅ็ไธไธๆใ็ธๅ็ๅๅ,็ฑๆไปฌไฝฟ็จ llama-perplexity ๆต้(BF16 ๆบๆไปถ = ๅ็
ง;BF16 ๆจกๅ็ PPL:en: 1.988,dev: 2.187,wiki: 8.688)ใๅ็
ง้ๅไธบ:bartowski(imatrix),ไปฅ็ธๅๆนๅผๆต้ใๆๅคงๅฐๆๅบใ
| ้ๅ | ๆฅๆบ | ๅคงๅฐ | KLD en | KLD dev | KLD wiki | PPL en | same top token (avg) |
|---|---|---|---|---|---|---|---|
Sakura-MiMo-V2.6-Distill-Qwen-9B-3.79GiB.gguf |
ๆฌไปๅบ | 3.79 GiB | 0.0642 | 0.0628 | 0.1146 | 2.065 | 89.2 % |
| IQ3_XXS | bartowski (imatrix) | 3.85 GiB | 0.0714 | 0.0662 | 0.1215 | 2.027 | 89.7 % |
Sakura-MiMo-V2.6-Distill-Qwen-9B-4.82GiB.gguf |
ๆฌไปๅบ | 4.82 GiB | 0.0179 | 0.0194 | 0.0285 | 2.020 | 94.3 % |
| IQ4_XS | bartowski (imatrix) | 4.87 GiB | 0.0198 | 0.0188 | 0.0290 | 2.013 | 94.6 % |
| Q4_K_M | bartowski (imatrix) | 5.44 GiB | 0.0127 | 0.0119 | 0.0200 | 2.007 | 95.5 % |
Sakura-MiMo-V2.6-Distill-Qwen-9B-5.55GiB.gguf |
ๆฌไปๅบ | 5.55 GiB | 0.0074 | 0.0076 | 0.0103 | 2.014 | 96.4 % |
| Q5_K_M | bartowski (imatrix) | 6.40 GiB | 0.0039 | 0.0034 | 0.0059 | 2.011 | 97.4 % |
ๆนๆณ:ๆฏไธชๆๆฌๅ 12 ไธชๅ,ๆฏๅ 512 ไธช token;ๆๆฌไธบ:่ฑๆ็ๅบๆๆฌใๅผๅ่ ๆๆฌ็ๅบ้ใ่ฑๆ็พ็ง็ๅบๆๆฌ(ๆช็จไบไปปไฝๆ กๅ)ใ่ฟๆฏๅจไธๅฐๆบๅจไธ็ๅๆฌกๆต้;็ธ่ฟๅคงๅฐไธ KLD ็็ปๅฐๅทฎๅผๅนถไธๆๆ่ดจ้ๆๅใ
ๅถไฝๆนๅผ
- ๅๅงๆจกๅ็ BF16 GGUF ๅ้่ฆๆง็ฉ้ตๆฅ่ช bartowski/MiMo-V2.6-Distill-Qwen-9B-GGUF;ๆไปฌๅๆ ทไฝฟ็จไบ่ฟไธค่ (ๆ่ฐขๅนถๅฝๅไบ bartowski)ใ
- ๆฏไธชๅคงๅๆ้็ฉ้ต้ๅฏนๆฏ็งๅ้็ฑปๅ(Q2_KใIQ2_SใIQ3_XXSใIQ3_SใQ3_KใIQ4_XSใQ4_KใQ5_KใQ6_KใQ8_0)ไฝฟ็จ
llama-quantizeๅ้่ฆๆง็ฉ้ตๅ้ๅไธๆฌก;ๆฏ็ง้ๆฉ็่ฏฏๅทฎๆ็ฉ้ตไผฐ่ฎก(ๆ้่ฆๆงๅ ๆ),ๅนถไนไปฅไธไธช ็จ็ๅฎ KL ๆฃๅบฆๆตๅพ็ๆๆๅบฆ็ณปๆฐ:ๆฏ็ปๅผ ้(ไพๅฆ FFN down ๆๅฝฑใๆณจๆๅ value ๆๅฝฑใ้ฆๅฐพๅ ๅฑ)่ขซๅ็ฌ็งปๅฐ่พไฝๆฏ็น็็ฑปๅ,ๅนถๅฐๅฎ้ ๆ็ KL ๆฃๅบฆไธๅ ถ้ๅ่ฏฏๅทฎ่ฟ่กๆฏ่พใ - ็ฒพ็กฎ็้ข็ฎๅ้ (ไปฅๅญ่ไธบๅไฝ็ๅค้่ๅ ้ฎ้ข)ไธบๆฏไธช็ฎๆ ๅคงๅฐไธบๆฏไธช็ฉ้ตๆ้ไธ็ง็ฑปๅ;ๆ็ปๆจกๅ็ฑๅทฒๅญๅจ็ๅผ ้็ป่ฃ ่ๆ,ๆ ้้ๆฐ้ๅใ ๆไปฌไธๅ ฌๅธ้ๅผ ้็้ๆฉใๅฝไธๅๅฑใๅฐๅผ ้ๅๅค token ้ขๆตๅฑไฟๆ้ซ็ฒพๅบฆใ
ๅฑ้
- ไป ็จ็ญ็็ๅบๆๆฌไธ็ KL ๆฃๅบฆๅๅฐๆๅบฆๆต้,ๆฒกๆไฝฟ็จไธๆธธๅบๅๆต่ฏ;่ฏทๅฟๅฐๅ ถ็่งฃไธบไปปๅก่ดจ้็ๅฃฐๆใ
- ้ๅๆปไผๆๅคฑไธไบ่ดจ้;ๆๅฐ็ๆไปถๆๅคฑๆๅคงใ
- ่ง่งๆๅฝฑๅจ
mmproj-MiMo-V2.6-Distill-Qwen-9B-f16.ggufๆฏ bartowski ็่ฝฌๆข็ๆฌ(ๆชๆนๅจ,F16);ๅพๅ้่ฆๅฎ,ๆไปฌๆฒกๆๅฏนๅฎ่ฟ่ก้ๅใ
่ด่ฐข
- ๅๅงๆจกๅ:MiMo-V2.6-Distill-Qwen-9B (Xiaomi MiMo, MIT),MIT ่ฎธๅฏ่ฏ(ไปฅ
LICENSE้้)ใ - BF16 GGUFใ้่ฆๆง็ฉ้ตๅๅ็ ง้ๅ:bartowskiใ
- GGUF ๆ ผๅผๅๅทฅๅ ท:ggml-org/llama.cpp (MIT)ใ
- Downloads last month
- 400
We're not able to determine the quantization variants.
Model tree for webmp3/Sakura-MiMo-V2.6-Distill-Qwen-9B-GGUF
Base model
Qwen/Qwen3.5-9B-Base