kani-tts-2-pt (LLM part) — GGUF

GGUF conversions of nineninesix/kani-tts-2-pt LLM part only.

Files

Converted with llama.cpp convert_hf_to_gguf.py and quantized with llama-quantize.
This checkpoint uses a custom HF architecture name (KaniTTS2ForCausalLM) but is compatible with the existing lfm2 GGUF arch in llama.cpp.
Non-LLM tensors (speaker conditioning projection) and unsupported learnable-RoPE parameters were omitted to keep this repo focused on the LLM part.

GGUF

Model size

68.6M params

Architecture

nemo_nano_codec

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

32-bit

Base model

Quantized

(2)

this model