| Multilingual powerhouse — testing for mobile deployment | 3morixd | open | 0 | disc |
| Qwen3-4B on phones — the Goldilocks model | 3morixd | open | 0 | disc |
| lm_head.weight is missing from weight_map in model.safetensors.index.json | zhaosiyuan | open | 0 | disc |
| main | BryanLane | draft | 0 | disc |
| 为什么我无法下载tokenizer.json（Why I can't download tokenizer.json） | Noahcbt | closed | 0 | disc |
| GIT_LFS_SKIP_SMUDGE=1 git clone git@hf.co:Qwen/Qwen3-4B | Ahmedkash | open | 0 | disc |
| Why is there a chat template for a base model? | dsouzaJithesh | open | 0 | disc |
| add AIBOM | fatima113 | open | 0 | disc |
| Add assistant mask support to Qwen3-4B | waleko | open | 0 | disc |
| UnslothVisionDataCollator problem | orkungedik | open | 0 | disc |
| Translation task in low-resource language can be done pretty well | luweigen | open | 0 | disc |
| Why are the new 4B and 8B models slower than the previous 7B-1M model?? | stev236 | closed | 0 | disc |
| Collections of Qwen3 4B model Bad Cases User Reviews and Comments | DeepNLP | open | 0 | disc |
| YaRN: is "performance" referring to quality or speed? | kmouratidis | open | 0 | disc |
| Use the more common reverse filter in template | tahayassine | open | 0 | disc |
| 【Evaluation】Best practice for evaluating Qwen3 !! | wangxingjun778 | open | 0 | disc |
| Add languages tag | de-francophones | open | 0 | disc |