Repository Issue Activity (beta)

#1308 C++/GGML based Fish Audio S2 Pro in audio.cpp. 3x faster than real time

#1325 Conversation.visualize exposes unsupported use_color and merge_audio_tokens parameters

#1291 Feature Request: Use FunASR for training data annotation

#1324 Verify evals on Papers with Code

#1305 s2.1pro

#1321 Improve Punjabi pronunciation through diacritic-aware Latin transliteration or G2P

#1320 Punjabi Gurmukhi input appears to mishandle addak and nasalization markers

#1310 Optimize native AR decoding by attending only to the active KV-cache prefix (1.97x compiled speedup on GV100)

#1263 Feature Request: Add language parameter to TTS API for cross-lingual CJK voice cloning

#1262 a way to run this model on DGX Spark GB10 since its basicly arm64

#1260 How to achieve voice consistency, without cloning.

#1272 Feature Request: OpenAI-compatible TTS endpoint for cloud API (api.fish.audio)

#1295 LoRA training cannot resume from checkpoint — Missing key(s) in state_dict for TextToSemantic

#1294 The Vietnamese language is being mispronounced (all models)

#1289 support for cuda 13

#1201 Best practices for Single / Multi-Speaker / Language Fine-tuning: Simultaneous training, Language Tags, and weighted rehearsal mixing?

#1266 'NoneType' object has no attribute 'encode' when using s1-mini in v.1.51

#1280 S2-Pro: bracket emotion control appears weak on emotion-neutral transcripts

#1273 No module named 'fish_speech.api'/fish_speech.infer/fish_speech

#1258 Cannot run on RTX 4090 with CUDA 13.2

#1168 Quantization or Optimization for S2-Pro Model

#1265 Web UI error

#1288 issue with s1 mini

#1212 6s-it speed on 3060 12GB

#1268 Cannot reproduce reported Seed-TTS-Eval results for s2-pro