Q4_k_m not available
Trying to download the q4 says is not available. Can you reupload it?
Edit:
I see now that all the rest beside the q2 and q3 is not available for download. ???
Hey @Trilogix1 ! All 9 files (Q2_K through BF16 split) are fully uploaded and present in the repo — I just verified server-side. Here are the confirmed sizes:
| Quant | Size |
|---|---|
| Q2_K | 135 GB |
| Q3_K_M | 177 GB |
| Q4_K_M | 224 GB |
| Q5_K_M | 262 GB |
| Q6_K | 303 GB |
| Q8_0 | 393 GB |
| BF16 (2 splits) | 739 GB |
If you're seeing "not available," it's most likely HuggingFace's CDN still propagating the large files. A few things to try:
- Clear browser cache and refresh the repo page
- Use the CLI instead:
huggingface-cli download timteh673/Qwen3.5-397B-A17B-Uncensored-GGUF Qwen3.5-397B-A17B-Uncensored-Q4_K_M.gguf - Wait 30-60 min and try again — HF can be slow to serve files >100GB after upload
Let me know if it's still giving you trouble after that!
Well, I keep getting unavailable. Can get it with Cli but why not with browser? Tried to clean cache and many browsers, weird.
Hey @Trilogix1 , that's a known HuggingFace issue with files >100GB — their web CDN sometimes marks them as "unavailable" even when the LFS objects are fully uploaded.
The CLI approach is the reliable workaround for now:
huggingface-cli download timteh673/Qwen3.5-397B-A17B-Uncensored-GGUF Qwen3.5-397B-A17B-Uncensored-Q4_K_M.gguf --local-dir ./
I've seen this affect other large model repos too (not just mine) — it seems to be a HF infrastructure limitation with their browser-based download handler for very large single files.
If you continue having issues with the CLI too, let me know and I'll investigate further. Also — a Stage 2 version (fine-tuned with Claude Opus reasoning traces) is being uploaded right now to a separate repo, so keep an eye out for that! ⚡
@timeteh673,
So far so good. You doing a good job :)