Skip to content

feat(qwen2_5vl): add model support and update related files#396

Merged
chenghuaWang merged 3 commits intoUbiquitousLearning:v2from
chenghuaWang:v2
Aug 25, 2025
Merged

feat(qwen2_5vl): add model support and update related files#396
chenghuaWang merged 3 commits intoUbiquitousLearning:v2from
chenghuaWang:v2

Conversation

@chenghuaWang
Copy link
Copy Markdown
Collaborator

feat(qwen2_5vl): add model support and update related files

  • Add Qwen2_5VL model configuration and implementation
  • Update main.cpp to use the new model for inference
  • Create new packages for TUI chat and model hub
  • Modify mllm-c API to use enum for return codes

- Add Qwen2_5VL model configuration and implementation
- Update main.cpp to use the new model for inference
- Create new packages for TUI chat and model hub
- Modify mllm-c API to use enum for return codes
…ecture

- Implement Gelu, Softmax, RMSNorm and Silu functions for X86 architecture
- Add corresponding kernels in mllm/backends/cpu/kernels/x86/
- Update GELUOp, QuickGELUOp, RMSNormOp, SiLUOp and SoftmaxOp to support X86
- Modify mllm/c.go to include new return codes and generation context
- Update mllm/mllm-c.hpp to define new C API functions for AR generation
- Add build configuration for X86 CLI in tasks/build_x86_cli.yaml
- Add Qwen2_5vl model support with new configuration and quantization files
- Implement naive broadcasting for AddOp in CPU operations
- Update Tensor class to include fromVector method
- Improve shape handling and broadcasting in abstract operations
- Modify CMakeLists to conditionally build C SDK binding
@chenghuaWang chenghuaWang merged commit 619aa19 into UbiquitousLearning:v2 Aug 25, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant