Chatterbox Turbo 350M Model#380
Merged
Merged
Conversation
fatchord
commented
Dec 15, 2025
fatchord
commented
Collaborator
- safetensors up on hf: https://huggingface.co/ResembleAI/chatterbox-turbo/tree/main
- Inference working with S3Gen (old)
- Inference working with S3Gen (distilled)
- Double-check older models work ok too
- New gradio app with events
manmay-nakhashi
approved these changes
Dec 15, 2025
Contributor
|
looks good |
LazyBusyYang
added a commit
to LazyBusyYang/chatterbox
that referenced
this pull request
May 7, 2026
…ltilingual, transformers 5.x) (#3) * Make russian-text-stresser optional (resemble-ai#376) * Update pyproject.toml * add automated install check * udpate to check master - not main * remove russian-stresser from dependency (resemble-ai#377) * Chatterbox Turbo 350M Model (resemble-ai#380) * add gpt2-medium config * setup turbo-specific hp * whitespace * t3 constructor logic * update gpt2 config values * fix t3 constructor * t3 safetensor loading correctly * everything loading ok * update from_pretrained method * inference working * new inference_turbo method * avoid tokenizer missing pad error * connect sampling params to .generate * add min_p to unused warning * inference with s3gen distilled * change to distilled model in from_local * norm loudness - relax punc norm * add gradio app * add silence to end of output wav * add events to gradio app * change default text * new dependencies * move advanced options to the righthand side of gradio app * remove russian-text-stresser and bump version * update readme * update readme * resize headings * links to sections * typo * update space link * chessy emojiis (might remove these later) * change to single emojii * remove most emojiis * Update README.md with new image * Update README.md remove extra char * add dedicated turbo example script --------- Co-authored-by: Zohaib Ahmed <zohaib@resemble.ai> * Fix ReadMe Banner Image (resemble-ai#381) * Add Podonos evaluation section and acknowledgement to README (resemble-ai#456) * Add Podonos evaluation links and acknowledgement * move evaluaton secion above acknowledgements * Update README.md (resemble-ai#441) fix bug in multilingual example * fix typo in README (resemble-ai#475) * fix: add map_location to torch.load for CPU/MPS support in multilingual model (resemble-ai#410) Add map_location parameter to torch.load() calls in ChatterboxMultilingualTTS to properly support CPU and MPS devices when loading CUDA-saved models. Changes: - Add map_location logic in from_local() for non-CUDA devices - Pass map_location to ve.pt and s3gen.pt loading - Pass map_location to Conditionals.load() for conds.pt - Add MPS availability check in from_pretrained() - Add str to torch.device conversion in Conditionals.load() Fixes resemble-ai#351, resemble-ai#357 Signed-off-by: majiayu000 <1835304752@qq.com> * Update Transformers + Gradio (resemble-ai#491) * update versions * fix token missing crash + wav path crash * fix crash in alignment stream analyzer * use latest perth watermark * Broaden dependency version ranges for Python 3.13+ compatibility (resemble-ai#495) * Add opt-in v3 multilingual checkpoint, skip analyzer for v3 (resemble-ai#516) * Add opt-in v3 multilingual checkpoint, skip analyzer for v3 * Remove alignment analyzer; lower rep_penalty default to 1.2; trim final speech token artifact * Refactor import statements in mtl_tts.py to remove duplicate import of SPEECH_VOCAB_SIZE --------- Signed-off-by: majiayu000 <1835304752@qq.com> Co-authored-by: Ollie McCarthy <fatchord@tutanota.com> Co-authored-by: manmay nakhashi <manmay.nakhashi@gmail.com> Co-authored-by: Zohaib Ahmed <zohaib@resemble.ai> Co-authored-by: Tedi Papajorgji <tedi.papajorgji@hotmail.com> Co-authored-by: Nicholas Loo <149620965+ResembleNick@users.noreply.github.com> Co-authored-by: Nicolas Müller <nicolas@resemble.ai> Co-authored-by: Matt <33789207+Kettukaa@users.noreply.github.com> Co-authored-by: lif <1835304752@qq.com> Co-authored-by: zihanjin428 <zihan@resemble.ai>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.