Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / OpenNMT/CTranslate2 issues and pull requests
#1790 - Missing converter : XLMRobertaFlashConfig
Issue -
State: open - Opened by ExtReMLapin 4 days ago
#1789 - Shallow Contextual Biasing for Whisper
Pull Request -
State: open - Opened by zwycl 5 days ago
#1788 - Using a partitioned A100 GPU via MIG with device_index and faster_index causing ctranslate2 error
Issue -
State: open - Opened by johnrisby 5 days ago
#1787 - Model is twice as large on first load
Issue -
State: open - Opened by winstxnhdw 6 days ago
#1786 - fix logits vocab
Pull Request -
State: closed - Opened by minhthuc2502 9 days ago
- 2 comments
#1785 - Mistral nemo
Pull Request -
State: closed - Opened by minhthuc2502 9 days ago
#1784 - Accept variable-length batch prompts for Whisper
Pull Request -
State: open - Opened by MahmoudAshraf97 10 days ago
- 15 comments
#1783 - Performance Regression in Whisper models when timestamp generation is enabled
Issue -
State: open - Opened by MahmoudAshraf97 10 days ago
- 2 comments
Labels: enhancement
#1782 - Python process crashes on exit under Windows with CUDA
Issue -
State: open - Opened by TechInterMezzo 11 days ago
- 1 comment
#1781 - Difference translation result after convert to ctranslate
Issue -
State: open - Opened by hieunguyenquoc 12 days ago
- 2 comments
#1780 - CUDNN 9 support
Issue -
State: open - Opened by AndrewMead10 17 days ago
- 1 comment
Labels: dependencies, python-release
#1779 - NO LOGITS RETURNS AFTER GENERATE
Issue -
State: open - Opened by LAnCeBabY 18 days ago
- 5 comments
#1778 - Wav2Vec2Bert ASR Inference Support
Pull Request -
State: closed - Opened by homink 18 days ago
- 6 comments
#1777 - correct docs regarding flash attention
Issue -
State: open - Opened by BBC-Esq 18 days ago
- 3 comments
Labels: documentation
#1776 - How to use 4-bit AWQ?
Issue -
State: open - Opened by BBC-Esq 18 days ago
- 9 comments
#1775 - Release 4.4.0 and flash attention with python [WIP]
Issue -
State: open - Opened by BBC-Esq 18 days ago
- 2 comments
#1774 - bump version 4.4.0
Pull Request -
State: closed - Opened by minhthuc2502 20 days ago
#1773 - Can I convert a ctranslate2 model to onnx?
Issue -
State: open - Opened by ashwingopinath 23 days ago
#1772 - support minimum gemma 2
Pull Request -
State: closed - Opened by minhthuc2502 23 days ago
#1771 - build failed on jetson agx orin (Error generating file: build/CMakeFiles/ctranslate2.dir/src/ops/flash-attention/./ctranslate2_generated_flash_fwd_split_hdim96_fp16_sm80.cu.o)
Issue -
State: open - Opened by cyu021 24 days ago
- 1 comment
#1770 - Bump actions/download-artifact from 3 to 4.1.7 in /.github/workflows
Pull Request -
State: open - Opened by dependabot[bot] 25 days ago
Labels: dependencies, github_actions
#1769 - Inference failed with "axis 2 has dimension xxxx but expected yyyy" error
Issue -
State: open - Opened by GangLiCN 26 days ago
- 2 comments
#1768 - How to early stop an encoding call?
Issue -
State: closed - Opened by mariano54 27 days ago
- 3 comments
#1767 - Error while converting to Ctranslate2 from openNMTPy.
Issue -
State: open - Opened by aryan1165 about 1 month ago
- 1 comment
#1766 - Reintroduce support for GPUs with Compute Capability 5.0
Pull Request -
State: open - Opened by giuliopaci about 1 month ago
#1765 - Reintroduce support for Compute Capability 5.0
Issue -
State: open - Opened by giuliopaci about 1 month ago
- 1 comment
#1764 - Failed to convert microsoft/Phi-3-medium-128k-instruct
Issue -
State: open - Opened by rbgo404 about 1 month ago
- 1 comment
#1763 - Support for DeepSeek models
Issue -
State: open - Opened by ByteForge786 about 1 month ago
#1762 - Convert Ctranslate2 model to Pytorch or TorchScript or PyTorch Lightning
Issue -
State: open - Opened by aryan1165 about 1 month ago
#1761 - Convert model.bin (fp32) to model.bin (int8)
Issue -
State: open - Opened by aryan1165 about 1 month ago
- 4 comments
#1760 - Bart models missing "embed_scale"
Pull Request -
State: open - Opened by l3utterfly about 1 month ago
- 1 comment
#1759 - When I run online for a long time, the gpu memory will get bigger and bigger
Issue -
State: open - Opened by dingjingzhen about 1 month ago
Labels: enhancement
#1758 - Wav2Vec2 upgrade with Conv1D options
Pull Request -
State: closed - Opened by homink about 2 months ago
- 3 comments
#1757 - set_random_seed does not make temperature based decoding deterministic
Issue -
State: open - Opened by ozancaglayan about 2 months ago
- 2 comments
#1756 - Support for ARM64 on Windows
Issue -
State: open - Opened by JanVlietinck about 2 months ago
- 3 comments
#1755 - Add log probs for all tokens
Pull Request -
State: closed - Opened by minhthuc2502 about 2 months ago
#1754 - Docker images not published
Issue -
State: closed - Opened by ales-t about 2 months ago
- 2 comments
#1753 - CI failing in the several recent PRs
Issue -
State: closed - Opened by homink about 2 months ago
- 2 comments
#1752 - Flash Attention regurgitates repeated tokens - seq2seq
Issue -
State: open - Opened by ArtanisTheOne about 2 months ago
- 1 comment
#1751 - fix: implement llama3 RoPE scaling type and fix converter
Pull Request -
State: closed - Opened by ebraraktas about 2 months ago
- 6 comments
#1750 - Converter falcon 2
Pull Request -
State: open - Opened by minhthuc2502 2 months ago
#1749 - feat: grouped conv1d
Pull Request -
State: closed - Opened by ebraraktas 2 months ago
- 9 comments
#1748 - IBM Power10 (VSX, MMA) support for ppc64le
Pull Request -
State: open - Opened by Dagamies 2 months ago
#1747 - Fix CI
Pull Request -
State: closed - Opened by minhthuc2502 2 months ago
#1746 - Bump torch from 2.1.0 to 2.2.0 in /python/tests
Pull Request -
State: closed - Opened by dependabot[bot] 2 months ago
Labels: dependencies
#1745 - Llama 3.1 support please?
Issue -
State: closed - Opened by BBC-Esq 2 months ago
- 2 comments
Labels: enhancement
#1744 - Converting to Llama architecture to expand models supported by Ctranslate2
Issue -
State: closed - Opened by BBC-Esq 2 months ago
- 1 comment
#1743 - Trouble converting Mistral-Nemo despite architecture indicating MistralForCausalLM
Issue -
State: open - Opened by BBC-Esq 2 months ago
- 8 comments
#1742 - The version of CUB in your include path is not compatible with this release of Thrust.
Issue -
State: closed - Opened by Tarinaky 2 months ago
- 1 comment
#1741 - Introduce better format for whisper models
Issue -
State: open - Opened by thewh1teagle 3 months ago
- 1 comment
#1740 - Other differences in the beam search implementation?
Issue -
State: open - Opened by koren-v 3 months ago
- 5 comments
#1739 - Implement python bindings for profiler
Pull Request -
State: open - Opened by yehudaorel 3 months ago
- 1 comment
#1738 - [WIP] HQQ Quant 4 bit
Pull Request -
State: open - Opened by minhthuc2502 3 months ago
#1737 - Falcon-11B support
Issue -
State: open - Opened by robertBrnnn 3 months ago
#1736 - [Feature Request] Expose Profiler to Python
Issue -
State: open - Opened by yehudaorel 3 months ago
- 1 comment
#1735 - Add gemma2 support
Issue -
State: open - Opened by Theodotus1243 3 months ago
- 1 comment
Labels: enhancement
#1734 - [Profiling] Help needed with Profiling CT2 on both CUDA and CPU
Issue -
State: open - Opened by prasad-pr-20 3 months ago
- 1 comment
#1733 - fix crash with return alternative on CUDA
Pull Request -
State: closed - Opened by minhthuc2502 3 months ago
#1732 - small improvement in flash attention
Pull Request -
State: closed - Opened by minhthuc2502 3 months ago
#1731 - Segmentation fault when using `return_alternatives` flag on NLLB model on cuda
Issue -
State: closed - Opened by Mukhammadsaid19 3 months ago
- 2 comments
#1730 - [feature request] Mixed quantizations.
Issue -
State: open - Opened by 0wwafa 3 months ago
- 1 comment
#1729 - Download ready to use model
Issue -
State: closed - Opened by thewh1teagle 3 months ago
- 3 comments
#1728 - Gemma model - help needed
Issue -
State: open - Opened by carolinaxxxxx 3 months ago
- 4 comments
#1727 - Quantzation AWQ GEMM + GEMV
Pull Request -
State: closed - Opened by minhthuc2502 3 months ago
- 5 comments
#1726 - Adding a layer to an existing model?
Issue -
State: closed - Opened by milot-mirdita 3 months ago
- 4 comments
#1725 - Feature Request: Support for CohereForAI AYA-23 models
Issue -
State: open - Opened by carolinaxxxxx 4 months ago
Labels: enhancement
#1724 - Inference for ctranslate2 using tensor parallel with mpi
Issue -
State: open - Opened by mohith56 4 months ago
- 1 comment
#1723 - fix pipline
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1723 - fix pipline
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1722 - bump version 4.3.1
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1722 - bump version 4.3.1
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1721 - Qwen2 Support?
Issue -
State: open - Opened by nguyenhoanganh2002 4 months ago
- 1 comment
Labels: enhancement
#1721 - Qwen2 Support?
Issue -
State: closed - Opened by nguyenhoanganh2002 4 months ago
- 2 comments
Labels: enhancement
#1720 - Does CT2/OpenNMT engine support Qualcomm SoC?
Issue -
State: open - Opened by Weisonorz 4 months ago
- 1 comment
#1720 - Does CT2/OpenNMT engine support Qualcomm SoC?
Issue -
State: open - Opened by Weisonorz 4 months ago
- 1 comment
#1719 - T5 inference result is all <pad>
Issue -
State: closed - Opened by k-washi 4 months ago
- 2 comments
#1718 - [*] improve log message for storage view content
Pull Request -
State: open - Opened by tirivo 4 months ago
- 1 comment
#1718 - [*] improve log message for storage view content
Pull Request -
State: open - Opened by tirivo 4 months ago
#1717 - Feature Request: Implement Static Cache and Quantization Techniques in CTranslate2
Issue -
State: open - Opened by Jiltseb 4 months ago
- 18 comments
Labels: enhancement
#1716 - Proper way to change alignment_heads and alignment_layer when using return_attention=True for converted HF Seq2Seq Transformers?
Issue -
State: open - Opened by jorirsan 4 months ago
#1715 - Facing issues with Ctranslate2 when working with Intel built-in GPU and oneDNN
Issue -
State: closed - Opened by prasad-pr-20 4 months ago
- 7 comments
#1714 - fix position bias in tensor parallel
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1713 - Dev/fix position bias in tp
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1712 - Ctranslate2 Pypi exceeds limit 20GB
Issue -
State: closed - Opened by minhthuc2502 4 months ago
- 1 comment
Labels: build, python-release
#1711 - Converter not working for NLLB models
Issue -
State: open - Opened by marcelgoya 4 months ago
- 9 comments
#1710 - Use specific CMAKE_CUDA_COMPILER instead of forcing CMake to use the same compiler for both CXX and CUDA
Pull Request -
State: open - Opened by alexlnkp 4 months ago
- 1 comment
#1709 - CUDA DeviceAllocate segfault
Issue -
State: open - Opened by drzraf 4 months ago
- 3 comments
#1708 - Different results when run with tensor parallelism
Issue -
State: open - Opened by subhalingamd 4 months ago
- 2 comments
#1707 - Support for Phi3-Small, Medium, and Vision
Issue -
State: open - Opened by BBC-Esq 4 months ago
- 1 comment
Labels: enhancement
#1706 - correctly determine INTEL_ROOT_DEFAULT using MKLROOT env variable
Pull Request -
State: closed - Opened by alexlnkp 4 months ago
- 7 comments
#1705 - use HINTS instead of PATHS
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1704 - Clang unusual switches wrongly hardcoded in resulting setup.py
Issue -
State: open - Opened by Manamama 4 months ago
- 3 comments
#1703 - Doesn't build without docker. libiomp5 not found
Issue -
State: closed - Opened by alexlnkp 4 months ago
- 6 comments
#1702 - Option --self_attn_type scaled-dot-flash is not supported (supported values are: scaled-dot)
Issue -
State: open - Opened by randomicity 4 months ago
- 6 comments
#1701 - bump version 4.3.0
Pull Request -
State: closed - Opened by minhthuc2502 4 months ago
#1700 - support phi-3 128k
Pull Request -
State: closed - Opened by minhthuc2502 5 months ago
#1699 - Whisper encode roughly 4x slower than openai/pytorch
Issue -
State: open - Opened by whispy-woods 5 months ago
- 1 comment
Labels: enhancement
#1698 - libctranslate2-81fc0d88.so.4.2.1 in python package has executable stack flag
Issue -
State: open - Opened by raivisdejus 5 months ago
Labels: enhancement
#1697 - CTranslate2 cmake error when trying to build the code from source with cuda support enabled on Windows.
Issue -
State: closed - Opened by prasad-pr-20 5 months ago
- 6 comments
#1696 - I got invalid conversion error when compile on linux
Issue -
State: closed - Opened by EyonJoshua91 5 months ago
- 2 comments