FlagRelease/MiniCPM-o-4.5-ascend-FlagOS
By FlagRelease · ↓ 15 · ♥ 0 · Updated 2026-03-09
What's in this repo
This page lists the files in FlagRelease/MiniCPM-o-4.5-ascend-FlagOS, including campplus.onnx, flow.decoder.estimator.fp32.onnx, speech_tokenizer_v2.onnx and more.
- 5 ONNX models (.onnx files)
- 375 files, including 145 .py, 63 .pyc, 48 .yaml, 23 .json, 16 .sh, 13 .md
- Last updated 2026-03-09
ONNX files
| File | Size | Download |
|---|---|---|
| assets/CosyVoice2-0.5B/campplus.onnx | — | Download |
| assets/CosyVoice2-0.5B/flow.decoder.estimator.fp32.onnx | — | Download |
| assets/CosyVoice2-0.5B/speech_tokenizer_v2.onnx | — | Download |
| assets/token2wav/campplus.onnx | — | Download |
| assets/token2wav/speech_tokenizer_v2_25hz.onnx | — | Download |
How to use
These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.
README
Introduction
FlagOS is a unified heterogeneous computing software stack for large models, co-developed with leading global chip manufacturers. With core technologies such as the FlagScale, together with vllm-plugin-fl, distributed training/inference framework, FlagGems universal operator library, FlagCX communication library, and FlagTree unified compiler, the FlagRelease platform leverages the FlagOS stack to automatically produce and release various combinations of . This enables efficient and automated model migration across diverse chips, opening a new chapter for large model deployment and application.
Based on this, the MiniCPM-o-4.5-ascend-FlagOS model is adapted for the Hygon chip using the FlagOS software stack, enabling:
Integrated Deployment
- Deep integration with the open-source FlagScale framework
- Out-of-the-box inference scripts with pre-configured hardware and software parameters
- Released FlagOS-Ascend container image supporting deployment within minutes
Consistency Validation
- Rigorously evaluated through benchmark testing: Performance and results from the FlagOS software stack are compared against native stacks on multiple public.
Technical Overview
FlagGems
FlagGems is a high-performance, generic operator library implemented in Triton language. It is built on a collection of backend-neutral kernels that aims to accelerate LLM (Large-Language Models) training and inference across diverse hardware platforms.
FlagTree
FlagTree is an open source, unified compiler for multiple AI chips project dedicated to developing a diverse ecosystem of AI chip compilers and related tooling platforms, thereby fostering and strengthening the upstream and downstream Triton ecosystem. Currently in its initial phase, the project aims to maintain compatibility with existing adaptation solutions while unifying the codebase to rapidly implement single-repository multi-backend support. For upstream model users, it provides unified compilation capabilities across multiple backends; for downstream chip manufacturers, it offers examples of Triton ecosystem integration.
FlagScale
FlagScale is a comprehensive toolkit designed to support the entire lifecycle of large models. It builds on the strengths of several prominent open-source projects, including Megatron-LM and vLLM, to provide a robust, end-to-end solution for managing and scaling large models.
FlagCX
FlagCX is a scalable and adaptive cross-chip communication library. It serves as a platform where developers, researchers, and AI engineers can collaborate on various projects, contribute to the development of cutting-edge AI solutions, and share their work with the global community.
FlagEval Evaluation Framework
FlagEval is a comprehensive evaluation system and open platform for large models launched in 2023. It aims to establish scientific, fair, and open benchmarks, methodologies, and tools to help researchers assess model and training algorithm performance. It features:
- Multi-dimensional Evaluation: Supports 800+ model evaluations across NLP, CV, Audio, and Multimodal fields, covering 20+ downstream tasks including language understanding and image-text generation.
- Industry-Grade Use Cases: Has completed horizontal evaluations of mainstream large models, providing authoritative benchmarks for chip-model performance validation.
Evaluation Results
In the following table, +x means FlagOS performs better than CUDA with x accuracy score, while -x means CUDA performs better. For FlagEval's benchmarks avg@1, +-5.00 means there's no significant difference between two AI-ecosystems.
Transformers version
Accuracy Difference between using FLAGOS and Ascend on Nvidia-CUDA
| Metrics | Difference with Nvidia-CUDA |
|---|---|
| Video-MME 0-shot avg@1 ↑ | 0.50% |
User Guide
Environment Setup
| Accelerator Card Driver Version | Driver Version: 6.3.13-v1.12.0a |
|---|---|
| CANN | Version: 25.2.0 |
| FlagGems | Version: 4.2.0 |
Operation Steps
Download FlagOS Image
docker pull harbor.baai.ac.cn/flagrelease-public/flagrelease-ascend-release-model_minicpm-o-45-tree_none-gems_4.2.0-scale_none-cx_none-python_3.11.13-torch_npu2.8.0-pcp_cann8.3.0.2.220_8.3.rc2-gpu_ascend001-arc_arm64-driver_25.2.0:latest
Download Open-source Model Weights
pip install modelscope
modelscope download --model FlagRelease/MiniCPM-o-4.5-ascend-FlagOS --local_dir /share/MiniCPMO45
Start the Container
#Container Startup
docker run -itd --name=ascend --privileged --ipc=host --network=host --shm-size=500G -w /workspace/cpmo -v /data/workspace-video-mme/video-mme:/data -v /usr/local/Ascend/driver:/usr/local/Ascend/driver -v /usr/local/Ascend/firmware:/usr/local/Ascend/firmware -v /usr/local/sbin/npu-smi:/usr/local/sbin/npu-smi -v /usr/local/dcmi:/usr/local/dcmi -v /usr/local/sbin:/usr/local/sbin -v /usr/bin/hostname:/usr/bin/hostname -v /etc/ascend_install.info:/etc/ascend_install.info -v /var/log/npu/:/usr/slog -v /etc/hccn.conf:/etc/hccn.conf -v /etc/localtime:/etc/localtime -v /etc/hosts:/etc/hosts -v /root/.cache:/root/.cache -v /share:/share harbor.baai.ac.cn/flagrelease-public/flagrelease-ascend-release-model_minicpm-o-45-tree_none-gems_4.2.0-scale_none-cx_none-python_3.11.13-torch_npu2.8.0-pcp_cann8.3.0.2.220_8.3.rc2-gpu_ascend001-arc_arm64-driver_25.2.0:latest
docker exec -it ascend bash
User Reference
MiniCPM-o-4.5 is a model that supports Omni-input and offers a wide range of usage scenarios. FlagRelease only provides migration results of the FlagOS software ecosystem across various chips and does not impose any restrictions on how users choose to utilize the model. You are encouraged to refer to the official OpenBMB README for flexible and comprehensive usage guidance. That said, we do provide a basic example: performing video-based question answering with spoken audio output.
cd /root/code
ln -s /share/MiniCPMO45 ./MiniCPMO45
python openbmb_main.py
Contributing
We warmly welcome global developers to join us:
- Submit Issues to report problems
- Create Pull Requests to contribute code
- Improve technical documentation
- Expand hardware adaptation support
License
The weight files are from MiniCPM-o-4.5 from OpenBMB, open source with apache2.0 licensehttps://www.apache.org/licenses/LICENSE-2.0.txt.
Model info
onnx
safetensors
minicpmo
custom_code
region:us
↓ 15
♥ 0
Created: 2026-01-30
Updated: 2026-03-09
View on Hugging FaceFiles (375)
.gitattributes
README.md
__pycache__/audio_utils.cpython-311.pyc
__pycache__/chunk_prefill_generate.cpython-311.pyc
__pycache__/configuration_minicpmo.cpython-311.pyc
__pycache__/configuration_minicpmtts.cpython-311.pyc
__pycache__/modeling_minicpmo.cpython-311.pyc
__pycache__/modeling_navit_siglip.cpython-311.pyc
__pycache__/processing_audio_minicpma.cpython-311.pyc
__pycache__/processing_image_minicpmv.cpython-311.pyc
__pycache__/processing_minicpmo.cpython-311.pyc
__pycache__/processing_streaming_mel.cpython-311.pyc
__pycache__/resampler.cpython-311.pyc
__pycache__/sliding_utils.cpython-311.pyc
__pycache__/tts_streaming_generate.cpython-311.pyc
__pycache__/utils.cpython-311.pyc
added_tokens.json
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/config.json
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/generation_config.json
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/merges.txt
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/model.safetensors
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/tokenizer_config.json
assets/CosyVoice2-0.5B/CosyVoice-BlankEN/vocab.json
assets/CosyVoice2-0.5B/README.md
assets/CosyVoice2-0.5B/asset/dingding.png
assets/CosyVoice2-0.5B/campplus.onnx
ONNX
assets/CosyVoice2-0.5B/config.json
assets/CosyVoice2-0.5B/configuration.json
assets/CosyVoice2-0.5B/cosyvoice2.yaml
assets/CosyVoice2-0.5B/flow.cache.pt
assets/CosyVoice2-0.5B/flow.decoder.estimator.fp32.onnx
ONNX
assets/CosyVoice2-0.5B/flow.encoder.fp16.zip
assets/CosyVoice2-0.5B/flow.encoder.fp32.zip
assets/CosyVoice2-0.5B/flow.pt
assets/CosyVoice2-0.5B/hift.pt
assets/CosyVoice2-0.5B/merges.txt
assets/CosyVoice2-0.5B/speech_tokenizer_v2.onnx
ONNX
assets/CosyVoice2-0.5B/tokenizer_config.json
assets/CosyVoice2-0.5B/vocab.json
assets/ascend/cosyvoice/.github/ISSUE_TEMPLATE/bug_report.md
assets/ascend/cosyvoice/.github/ISSUE_TEMPLATE/feature_request.md
assets/ascend/cosyvoice/.github/workflows/lint.yml
assets/ascend/cosyvoice/.github/workflows/stale-issues.yml
assets/ascend/cosyvoice/.gitignore
assets/ascend/cosyvoice/.gitmodules
assets/ascend/cosyvoice/CODE_OF_CONDUCT.md
assets/ascend/cosyvoice/FAQ.md
assets/ascend/cosyvoice/LICENSE
assets/ascend/cosyvoice/README.md
assets/ascend/cosyvoice/asset/cross_lingual_prompt.wav
assets/ascend/cosyvoice/asset/dingding.png
assets/ascend/cosyvoice/asset/zero_shot_prompt.wav
assets/ascend/cosyvoice/cosyvoice/__init__.py
assets/ascend/cosyvoice/cosyvoice/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/bin/average_model.py
assets/ascend/cosyvoice/cosyvoice/bin/export_jit.py
assets/ascend/cosyvoice/cosyvoice/bin/export_onnx.py
assets/ascend/cosyvoice/cosyvoice/bin/train.py
assets/ascend/cosyvoice/cosyvoice/cli/__init__.py
assets/ascend/cosyvoice/cosyvoice/cli/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/cli/__pycache__/cosyvoice.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/cli/__pycache__/frontend.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/cli/__pycache__/model.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/cli/cosyvoice.py
assets/ascend/cosyvoice/cosyvoice/cli/frontend.py
assets/ascend/cosyvoice/cosyvoice/cli/model.py
assets/ascend/cosyvoice/cosyvoice/dataset/__init__.py
assets/ascend/cosyvoice/cosyvoice/dataset/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/dataset/__pycache__/processor.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/dataset/dataset.py
assets/ascend/cosyvoice/cosyvoice/dataset/processor.py
assets/ascend/cosyvoice/cosyvoice/flow/DiT/dit.py
assets/ascend/cosyvoice/cosyvoice/flow/DiT/modules.py
assets/ascend/cosyvoice/cosyvoice/flow/__pycache__/decoder.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/flow/__pycache__/flow.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/flow/__pycache__/flow_matching.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/flow/decoder.py
assets/ascend/cosyvoice/cosyvoice/flow/flow.py
assets/ascend/cosyvoice/cosyvoice/flow/flow_matching.py
assets/ascend/cosyvoice/cosyvoice/flow/length_regulator.py
assets/ascend/cosyvoice/cosyvoice/hifigan/__pycache__/discriminator.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/hifigan/__pycache__/f0_predictor.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/hifigan/__pycache__/generator.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/hifigan/__pycache__/hifigan.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/hifigan/discriminator.py
assets/ascend/cosyvoice/cosyvoice/hifigan/f0_predictor.py
assets/ascend/cosyvoice/cosyvoice/hifigan/generator.py
assets/ascend/cosyvoice/cosyvoice/hifigan/hifigan.py
assets/ascend/cosyvoice/cosyvoice/llm/__pycache__/llm.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/llm/llm.py
assets/ascend/cosyvoice/cosyvoice/tokenizer/__pycache__/tokenizer.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/tokenizer/assets/multilingual_zh_ja_yue_char_del.tiktoken
assets/ascend/cosyvoice/cosyvoice/tokenizer/tokenizer.py
assets/ascend/cosyvoice/cosyvoice/transformer/__init__.py
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/activation.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/attention.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/convolution.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/embedding.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/encoder_layer.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/label_smoothing_loss.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/positionwise_feed_forward.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/subsampling.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/__pycache__/upsample_encoder.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/transformer/activation.py
assets/ascend/cosyvoice/cosyvoice/transformer/attention.py
assets/ascend/cosyvoice/cosyvoice/transformer/convolution.py
assets/ascend/cosyvoice/cosyvoice/transformer/decoder.py
assets/ascend/cosyvoice/cosyvoice/transformer/decoder_layer.py
assets/ascend/cosyvoice/cosyvoice/transformer/embedding.py
assets/ascend/cosyvoice/cosyvoice/transformer/encoder.py
assets/ascend/cosyvoice/cosyvoice/transformer/encoder_layer.py
assets/ascend/cosyvoice/cosyvoice/transformer/label_smoothing_loss.py
assets/ascend/cosyvoice/cosyvoice/transformer/positionwise_feed_forward.py
assets/ascend/cosyvoice/cosyvoice/transformer/subsampling.py
assets/ascend/cosyvoice/cosyvoice/transformer/upsample_encoder.py
assets/ascend/cosyvoice/cosyvoice/utils/__init__.py
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/class_utils.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/common.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/file_utils.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/frontend_utils.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/losses.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/__pycache__/mask.cpython-311.pyc
assets/ascend/cosyvoice/cosyvoice/utils/class_utils.py
assets/ascend/cosyvoice/cosyvoice/utils/common.py
assets/ascend/cosyvoice/cosyvoice/utils/executor.py
assets/ascend/cosyvoice/cosyvoice/utils/file_utils.py
assets/ascend/cosyvoice/cosyvoice/utils/frontend_utils.py
assets/ascend/cosyvoice/cosyvoice/utils/losses.py
assets/ascend/cosyvoice/cosyvoice/utils/mask.py
assets/ascend/cosyvoice/cosyvoice/utils/scheduler.py
assets/ascend/cosyvoice/cosyvoice/utils/train_utils.py
assets/ascend/cosyvoice/cosyvoice/vllm/cosyvoice2.py
assets/ascend/cosyvoice/docker/Dockerfile
assets/ascend/cosyvoice/example.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/Dockerfile
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/README.md
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/huggingface_to_pretrained.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/infer_dataset.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/prepare_data.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/pretrained_to_huggingface.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/requirements.txt
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/reward_tts.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/run.sh
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/scripts/compute_wer.sh
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/scripts/offline-decode-files.py
assets/ascend/cosyvoice/examples/grpo/cosyvoice2/token2wav_asr_server.py
assets/ascend/cosyvoice/examples/libritts/cosyvoice/conf/cosyvoice.yaml
assets/ascend/cosyvoice/examples/libritts/cosyvoice/conf/ds_stage2.json
assets/ascend/cosyvoice/examples/libritts/cosyvoice/local/download_and_untar.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice/local/prepare_data.py
assets/ascend/cosyvoice/examples/libritts/cosyvoice/local/prepare_reject_sample.py
assets/ascend/cosyvoice/examples/libritts/cosyvoice/path.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice/run.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice/tts_text.json
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/conf/cosyvoice2.yaml
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/conf/ds_stage2.json
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/path.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/run.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/run_dpo.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice2/tts_text.json
assets/ascend/cosyvoice/examples/libritts/cosyvoice3/conf/cosyvoice3.yaml
assets/ascend/cosyvoice/examples/libritts/cosyvoice3/conf/ds_stage2.json
assets/ascend/cosyvoice/examples/libritts/cosyvoice3/path.sh
assets/ascend/cosyvoice/examples/libritts/cosyvoice3/run.sh
assets/ascend/cosyvoice/examples/magicdata-read/cosyvoice/local/download_and_untar.sh
assets/ascend/cosyvoice/examples/magicdata-read/cosyvoice/local/prepare_data.py
assets/ascend/cosyvoice/examples/magicdata-read/cosyvoice/path.sh
assets/ascend/cosyvoice/examples/magicdata-read/cosyvoice/run.sh
assets/ascend/cosyvoice/examples/magicdata-read/cosyvoice/tts_text.json
assets/ascend/cosyvoice/requirements.txt
assets/ascend/cosyvoice/runtime/python/Dockerfile
assets/ascend/cosyvoice/runtime/python/fastapi/client.py
assets/ascend/cosyvoice/runtime/python/fastapi/server.py
assets/ascend/cosyvoice/runtime/python/grpc/client.py
assets/ascend/cosyvoice/runtime/python/grpc/cosyvoice.proto
assets/ascend/cosyvoice/runtime/python/grpc/server.py
assets/ascend/cosyvoice/runtime/triton_trtllm/Dockerfile.server
assets/ascend/cosyvoice/runtime/triton_trtllm/README.DIT.md
assets/ascend/cosyvoice/runtime/triton_trtllm/README.md
assets/ascend/cosyvoice/runtime/triton_trtllm/client_grpc.py
assets/ascend/cosyvoice/runtime/triton_trtllm/client_http.py
assets/ascend/cosyvoice/runtime/triton_trtllm/docker-compose.dit.yml
assets/ascend/cosyvoice/runtime/triton_trtllm/docker-compose.yml
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/audio_tokenizer/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/audio_tokenizer/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/cosyvoice2/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/cosyvoice2/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/cosyvoice2_dit/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/cosyvoice2_dit/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/speaker_embedding/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/speaker_embedding/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/tensorrt_llm/1/.gitkeep
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/tensorrt_llm/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/token2wav/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/token2wav/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/token2wav_dit/1/model.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/token2wav_dit/1/token2wav_dit.py
assets/ascend/cosyvoice/runtime/triton_trtllm/model_repo/token2wav_dit/config.pbtxt
assets/ascend/cosyvoice/runtime/triton_trtllm/offline_inference.py
assets/ascend/cosyvoice/runtime/triton_trtllm/requirements.txt
assets/ascend/cosyvoice/runtime/triton_trtllm/run.sh
assets/ascend/cosyvoice/runtime/triton_trtllm/run_stepaudio2_dit_token2wav.sh
assets/ascend/cosyvoice/runtime/triton_trtllm/scripts/convert_checkpoint.py
assets/ascend/cosyvoice/runtime/triton_trtllm/scripts/fill_template.py
assets/ascend/cosyvoice/runtime/triton_trtllm/scripts/test_llm.py
assets/ascend/cosyvoice/runtime/triton_trtllm/streaming_inference.py
assets/ascend/cosyvoice/runtime/triton_trtllm/token2wav.py
assets/ascend/cosyvoice/runtime/triton_trtllm/token2wav_dit.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/.env.example
assets/ascend/cosyvoice/third_party/Matcha-TTS/.github/PULL_REQUEST_TEMPLATE.md
assets/ascend/cosyvoice/third_party/Matcha-TTS/.github/codecov.yml
assets/ascend/cosyvoice/third_party/Matcha-TTS/.github/dependabot.yml
assets/ascend/cosyvoice/third_party/Matcha-TTS/.github/release-drafter.yml
assets/ascend/cosyvoice/third_party/Matcha-TTS/.gitignore
assets/ascend/cosyvoice/third_party/Matcha-TTS/.pre-commit-config.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/.project-root
assets/ascend/cosyvoice/third_party/Matcha-TTS/.pylintrc
assets/ascend/cosyvoice/third_party/Matcha-TTS/LICENSE
assets/ascend/cosyvoice/third_party/Matcha-TTS/MANIFEST.in
assets/ascend/cosyvoice/third_party/Matcha-TTS/Makefile
assets/ascend/cosyvoice/third_party/Matcha-TTS/README.md
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/callbacks/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/callbacks/model_checkpoint.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/callbacks/model_summary.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/callbacks/none.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/callbacks/rich_progress_bar.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/data/hi-fi_en-US_female.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/data/ljspeech.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/data/vctk.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/debug/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/debug/fdr.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/debug/limit.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/debug/overfit.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/debug/profiler.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/eval.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/experiment/hifi_dataset_piper_phonemizer.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/experiment/ljspeech.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/experiment/ljspeech_min_memory.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/experiment/multispeaker.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/extras/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/hparams_search/mnist_optuna.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/hydra/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/local/.gitkeep
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/aim.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/comet.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/csv.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/many_loggers.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/mlflow.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/neptune.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/tensorboard.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/logger/wandb.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/model/cfm/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/model/decoder/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/model/encoder/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/model/matcha.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/model/optimizer/adam.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/paths/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/train.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/cpu.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/ddp.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/ddp_sim.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/default.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/gpu.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/configs/trainer/mps.yaml
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/VERSION
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/app.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/cli.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/data/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/data/components/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/data/text_mel_datamodule.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/LICENSE
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/README.md
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/__pycache__/models.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/__pycache__/xutils.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/config.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/denoiser.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/env.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/meldataset.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/models.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/hifigan/xutils.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/baselightningmodule.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/__pycache__/decoder.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/__pycache__/flow_matching.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/__pycache__/transformer.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/decoder.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/flow_matching.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/text_encoder.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/components/transformer.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/models/matcha_tts.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/onnx/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/onnx/export.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/onnx/infer.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/text/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/text/cleaners.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/text/numbers.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/text/symbols.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/train.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/__init__.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/audio.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/instantiators.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/logging_utils.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/pylogger.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/rich_utils.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/__pycache__/utils.cpython-311.pyc
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/audio.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/generate_data_statistics.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/instantiators.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/logging_utils.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/model.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/monotonic_align/__init__.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/monotonic_align/core.pyx
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/monotonic_align/setup.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/pylogger.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/rich_utils.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/matcha/utils/utils.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/notebooks/.gitkeep
assets/ascend/cosyvoice/third_party/Matcha-TTS/pyproject.toml
assets/ascend/cosyvoice/third_party/Matcha-TTS/requirements.txt
assets/ascend/cosyvoice/third_party/Matcha-TTS/scripts/schedule.sh
assets/ascend/cosyvoice/third_party/Matcha-TTS/setup.py
assets/ascend/cosyvoice/third_party/Matcha-TTS/synthesis.ipynb
assets/ascend/cosyvoice/tools/extract_embedding.py
assets/ascend/cosyvoice/tools/extract_speech_token.py
assets/ascend/cosyvoice/tools/make_parquet_list.py
assets/ascend/cosyvoice/vllm_example.py
assets/ascend/cosyvoice/webui.py
assets/ascend/weights/flow_linux_aarch64.om
assets/ascend/weights/flow_static.om
assets/ascend/weights/speech_linux_aarch64.om
assets/token2wav/campplus.onnx
ONNX
assets/token2wav/flow.pt
assets/token2wav/flow.yaml
assets/token2wav/hift.pt
assets/token2wav/speech_tokenizer_v2_25hz.onnx
ONNX
audio_utils.py
chunk_prefill_generate.py
config.json
configuration_minicpmo.py
configuration_minicpmtts.py
duplex_config.json
duplex_utils.py
generation_config.json
main.py
merges.txt
model.safetensors
modeling_minicpmo.py
modeling_navit_siglip.py
preprocessor_config.json
processing_audio_minicpma.py
processing_image_minicpmv.py
processing_minicpmo.py
processing_streaming_mel.py
resampler.py
sliding_utils.py
special_tokens_map.json
stream_decoder.py
stream_providers.py
tokenization_minicpmo_fast.py
tokenizer.json
tokenizer_config.json
tts_streaming_generate.py
utils.py
vocab.json
Related models
More from FlagRelease
More models
No task type declared.