-
Trim redundant note from the MTP section
-
Remove stray backup files from the checkpoint
-
Tag as image-text-to-text; document vision and video support
-
license_name: qwen-community-1.0, matching upstream
-
Fix license: Qwen Community License 1.0, not Apache-2.0
-
Add Qwen Community License 1.0 text
-
library_name: vllm; explain the root config.json
-
Add root config.json so the Hub can count downloads
-
Lead with the lean download: skip 99 GB of never-read shards
-
Add NVFP4 checkpoint (131 shards)
-
Credits: add Opus 5
-
Add quantized PLE n-gram sidecar (128 shards)
-
Drop primitive-ai's sample shard and README; full sidecar ships separately
-
Add vLLM PLE overlays (primitive-ai, Apache-2.0)
-
Add model card
-
initial commit
