Image-Text-to-Text
Transformers
Safetensors
English
qwen3_5_text
text-generation
dag-reasoning
valiant
valiant-labs
qwen
qwen-3.5
qwen-3.5-27b
27b
reasoning
directed-acyclic-graph
graph
logic
analysis
programming
knowledge
root-cause-analysis
economics
business
business-management
finance
law
supply-chain
logistics
software-engineering
cybersecurity
architecture
energy
politics
problem-solving
creative
analytical
expert
rationality
conversational
chat
instruct
Upload folder using huggingface_hub
#1
by sequelbox - opened
No description provided.
This PR merges the untrained 'visual' and 'mtp' base weights from Qwen/Qwen3.5-27B. This does not modify the trained weights, but improves the user experience by directly 'unlocking' the original Qwen 3.5 multimodal capabilities instead of forcing the user to swap or merge them.
We're not expecting any additional changes to the weights from this point, but if anything is found it will be addressed.
sequelbox changed pull request status to merged