Views
No views yet
| Run locally | Pool multiple machines | OpenAI-compatible | Package variant |
|---|---|---|---|
| Private inference on your hardware | Split layers across peers | Serve /v1/chat/completions locally | Q4_K_M layer package |
| Property | Value |
|---|---|
| Source model | Qwen/Qwen3-VL-8B-Instruct-GGUF |
| Model id | Qwen/Qwen3-VL-8B-Instruct-GGUF:Q4_K_M |
| Family | Qwen3 |
| Parameter scale | 8B |
| Quantization | Q4_K_M |
| Layer count | 36 |
| Activation width | 4096 |
| Package size | 6.7 GB |
| Source file | Qwen3VL-8B-Instruct-Q4_K_M.gguf |
| Package repo | namepd/Qwen3-VL-8B-Instruct-Q4_K_M-layers |
| License | apache-2.0 from Qwen/Qwen3-VL-8B-Instruct-GGUF |
1# Run this on each machine that should contribute memory/compute.
2mesh-llm serve --model "namepd/Qwen3-VL-8B-Instruct-Q4_K_M-layers" --split1# Check the mesh and discover the OpenAI-compatible model name.
2curl -s http://localhost:3131/api/status
3curl -s http://localhost:3131/v1/models1# Send an OpenAI-compatible chat request.
2curl -s http://localhost:3131/v1/chat/completions \
3 -H "Content-Type: application/json" \
4 -d '{
5 "model": "Qwen/Qwen3-VL-8B-Instruct-GGUF:Q4_K_M",
6 "messages": [{"role": "user", "content": "Write a tiny hello-world function in Rust."}],
7 "max_tokens": 128
8 }'| Property | Value |
|---|---|
| Format | layer-package |
| Canonical source ref | Qwen/Qwen3-VL-8B-Instruct-GGUF@f982a07559d4a2f6c8744d840bf6fccab30eea96/Qwen3VL-8B-Instruct-Q4_K_M.gguf |
| Source revision | f982a07559d4a2f6c8744d840bf6fccab30eea96 |
| Source SHA-256 | 67d1659bfe71b89d50b45a4ad1a9e5b997e5bb16ce5da66a6a6167abd569e9e2 |
| Skippy ABI | 0.1.38 |
| Package manifest SHA-256 | dd6e6a120672480455a98eab01b05ef2994db9f68f53df050f3f9fb863dd011c |
| Artifact | Path | Contents | SHA-256 |
|---|---|---|---|
| Manifest | model-package.json | Package schema, source identity, checksums | dd6e6a120672480455a98eab01b05ef2994db9f68f53df050f3f9fb863dd011c |
| Metadata | shared/metadata.gguf | 0 tensors, 5.7 MB | 9feef8d599a3ede981d6254ed23f45191439f4c3e9f8fda763d9827bcac76b27 |
| Embeddings | shared/embeddings.gguf | 1 tensors, 339.5 MB | df24139105af8289aa14a281dfdc6758c9fbab38b3e6d21749ae39d5e2a42c93 |
| Output head | shared/output.gguf | 2 tensors, 492.5 MB | a8fcd68146baac9ca49efea4d89d6eb59dbf21343988f4fa80402ee42a0af974 |
| Transformer layers | layers/layer-*.gguf | 36 layer artifacts, 396 tensors, 4.1 GB | see model-package.json |
| Projector | projectors/mmproj-Qwen3VL-8B-Instruct-F16.gguf | mmproj projector, 1.1 GB | ca524100ebf825c9a870db1c580d03879e0da0ab2541697e2458e64891cf9d38 |
| Projector | projectors/mmproj-Qwen3VL-8B-Instruct-Q8_0.gguf | mmproj projector, 717.4 MB | c6ba85508d82f42590e6eb77d5340369ab6fecf107a7561d809523d8aa5f3bfd |
mesh-llm ref main.
Each artifact is checksummed as it is written, uploaded to this repository, and removed from the job workspace before the next artifact is produced.skippy-model-package write-package "/source/Qwen3VL-8B-Instruct-Q4_K_M.gguf" --out-dir "/tmp/meshllm-layer-job-namepd_Qwen3-VL-8B-Instruct-Q4_K_M-layers-232/package"