mradermacher

/

model_requests

like 171

It's queued!

You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking-uncensored-heretic-GGUF for quants to appear.

Is everything all right? Can I hope for the quants? I noticed that the model disappeared from the list on http://hf.tst.eu/status.html yesterday. Sorry — this is my first request, so I got a bit worried :)

error loading model: missing tensor 'blk.96.attn_norm.weight'

@llmfan46 is model architecture modified ? very weird that it doesnt work, we have quite a few models that fail in similar way past few days 🤔

error loading model: missing tensor 'blk.96.attn_norm.weight'

@llmfan46 is model architecture modified ? very weird that it doesnt work, we have quite a few models that fail in similar way past few days 🤔

Could this be due to a transformers compatibility issue with older version?

Or maybe it could be related to the fact that newest versions of llama.cpp does expect MTPs from Qwen3.5/3.6 models and tries to extract them even when they are not there? If that is the issue, you need to add --no-mtp, like this:

python convert_hf_to_gguf.py C:\Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking-uncensored-heretic --outfile C:\Qwen3.5-40B-Claude-4.5-Opus-High-Reasoning-Thinking-uncensored-heretic-BF16.gguf --outtype bf16 --no-mtp

@RichardErkhov Did the solution proposed by llmfan46 help? Is there any chance of getting the quants?

Sorry for any mistakes — English isn’t my native language.