pipecat-ai/phonellm-alpha-1
Full fine-tune vs nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16. Fine-tune: 97% of the weights changed, across 52 of 52 layers.
Tensors changed: other 175/207, Embeddings 0/1, Norms 46/76, MLP 2680/2691, Attention 24/24, lm_head 1/1
4 derivatives checked
- EryriLabs/phonellm-alpha-1-GGUF: Same weights, converted
- mradermacher/phonellm-alpha-1-i1-GGUF: Quantization
- pipecat-ai/phonellm-alpha-1-nvfp4: Quantization
- mradermacher/phonellm-alpha-1-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.