39 MB

Ctrl+K

8 contributors

History: 47 commits

Marcos

Claude

feat: Implementar treinamento Q-Former com Common Voice para compreensão de áudio

21fe3b9 about 1 year ago

.claude
refactor: clean project structure to essentials only about 1 year ago
llama_omni2
refactor: Remove CosyVoice e atualiza para TTS externos about 1 year ago
models_download
refactor: Remove CosyVoice e atualiza para TTS externos about 1 year ago
pipelines
feat: Setup completo para treinamento Qwen3-0.6B speech embeddings about 1 year ago
tests
feat: Treinamento Qwen3-0.6B Stage I com LoRA + Scripts de teste about 1 year ago
training
feat: Implementar treinamento Q-Former com Common Voice para compreensão de áudio about 1 year ago
.gitignore

1.93 kB
refactor: Move models from models/vendor/ to models/ about 1 year ago
.install_status

10 Bytes
feat: Treinamento Qwen3-0.6B Stage I com LoRA + Scripts de teste about 1 year ago
CLAUDE.md

14.1 kB
feat: Solução definitiva para português - Qwen3-8B sem modo thinking about 1 year ago
README.md

915 Bytes
docs: add essential README.md about 1 year ago
RELATORIO_FINAL.md

4.58 kB
feat: LLaMA-Omni2 pipeline funcionando com GPU about 1 year ago
VERSION

91 Bytes
feat: Adiciona scripts de teste e verificação da instalação about 1 year ago
check_install.sh

1.95 kB
feat: Install script completamente em background about 1 year ago
install.sh

4.12 kB
feat: Install script completamente em background about 1 year ago
llama-omni2-paper.pdf

741 kB
feat: Solução definitiva - enable_thinking=False elimina thinking mode about 1 year ago
pyproject.toml

1.62 kB
Initial commit - LLaMA-Omni2 code only about 1 year ago
requirements.txt

885 Bytes
Add requirements.txt and installation guide about 1 year ago
server.py

12.8 kB
refactor: clean project structure to essentials only about 1 year ago
start.sh

1.88 kB
refactor: clean project structure to essentials only about 1 year ago