mradermacher

/

model_requests

like 132

Hi, it seems to have a broken config file
taozi555/NVIDIA-Nemotron-3-Nano-30B-A3B-novelrp-BF16: no architectures entry (malformed JSON string, neither tag, array, object, number, string or atom, at character offset 1630 (before "Infinity\n ],\n "t...")
I will try to force queue it anyways, but not sure if it will work

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#NVIDIA-Nemotron-3-Nano-30B-A3B-novelrp-BF16-GGUF for quants to appear.

We boycott all recently released Nemotron based models for this exact reason. NVIDIA purposefully put invalid JSON in their model config despite being fully aware that doing so causes downstream issues for everyone. We even informed them about this last summer and they agree that their model config contains invalid JSON and does NOT follow the JSON specification yet they decided not to fix it so as response we decided to no longer quant their models as it’s a massive pain for us to manually fix every single of their broken config files before quantizing. There is nothing that stops us from fixing the config file and manually quant it anyways if this model is very important to you.

Tks a lot for the response and the effort. I'm a big fan of role play but i have no good hardware. This model runs fast on my CPU and i continually looking for RP fine tunnings to play with.
(I see you already created the GGUF. I'll try it and let you know if it worth the effort.
Tks again.