Re-convert with current codec.cpp: c0_head->heads_0 (fixes codec_lm_create missing-tensor)

Re-convert with current codec.cpp: c0_head->heads_0 (fixes codec_lm_create missing-tensor)

Re-convert with current codec.cpp: c0_head->heads_0 (fixes codec_lm_create missing-tensor)

Re-convert with current codec.cpp: c0_head->heads_0 (fixes codec_lm_create missing-tensor)

Re-convert with current codec.cpp: c0_head->heads_0 (fixes codec_lm_create missing-tensor)

feat: T3 backbone q3/q2 quants

feat: T3 backbone q5/q4 quants

feat: T3 backbone f32/f16/bf16/q8/q6 (Llama-520M arch)

feat: native codec_lm layout (codec.gguf bundles cond_enc + text_emb + speech_pos_emb)

Delete files codec-f16.gguf codec-f32.gguf codec-q4_k_m.gguf codec-q5_k_m.gguf codec-q8_0.gguf mtl_tokenizer.json s3t.gguf t3-extras.gguf t3-f16.gguf t3-f32.gguf t3-q2_k.gguf t3-q3_k_l.gguf t3-q3_k_m.gguf t3-q3_k_s.gguf t3-q4_0.gguf t3-q4_1.gguf t3-q4_k_m.gguf t3-q4_k_s.gguf t3-q5_0.gguf t3-q5_1.gguf t3-q5_k_m.gguf t3-q5_k_s.gguf t3-q6_k.gguf t3-q8_0.gguf with huggingface_hub

Delete files upload.log quants.log with huggingface_hub

feat: Chatterbox-Multilingual T3 GGUF (23-lang) + shared S3G/S3T

initial commit