Problemsby IbisTrouble with Hugging Face size
MLOps3 posts from 3 people2d active+22 posts in the last 7 days, 1 the 7 days before (steady)
Posts per day
The posts behind it
3, newest firstSmolVLM forward with images fails in v5.20.1: shape '[-1, 14, ...]' is invalid (inputs_merger uses vision patch_size)### System Info - `transformers` version: 5.20.1 (also current main, c05860687f) - Platform: Linux - Python version: 3.10 - PyTorch version: 2.13.0 (CPU and CUDA) ### Who can help? @Arnav66692 (author of #49557) @zucchini-nlp ### Information - [ ] The official example scripts -...huggingface/transformers · albertvillanova1 repliesOpen ↗[Bug] Qwen3-MoE router auxiliary loss is multiplied by the DDP world size in Trainer### Follow-up to #48690 Re-filing #48690 after its automatic stale closure on 2026-10-19. The original author account cannot reopen that issue. The proposed fixes in #48691 and #48706 are both still open and unmerged as of 2026-10-19. If maintainers prefer to continue in the...huggingface/transformers · gss102820251 repliesOpen ↗Inkling-Small: width of the routed experts resolves to 3072 instead of 2048 (moe_intermediate_size is never read from the checkpoint's config)...2048` to `text_config` in the model repository's `config.json`. Workaround we use: write `moe_intermediate_size: 2048` into `text_config` before loading. With that, a copy of the model split by layers across several machines gives the same next token as the served NVFP4...huggingface/transformers · symbai-tech3 repliesOpen ↗
Companies and products named
About this problem
Evidence
3 posts from 3 people in 1 place, about 2 a week over 12 days. Mostly on huggingface/transformers. Tools named alongside: PyTorch, GitHub, NVIDIA.
Frustration Frustration 0 of 3· Seen on GitHub issues
How it was grouped
Posts about Hugging Face with no known issue, sharing the word "size". First post Oct 8, 2026, latest Oct 19, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
Other Hugging Face problems
Fine-tuning Hugging Face is costly and unpredictableTrouble with Hugging Face generateHugging Face users keep filing bugsTrouble with Hugging Face cruiseRunning Hugging Face locally hits memory and speed limitsHugging Face upgrades break working setupsTrouble with Hugging Face docsTrouble with Hugging Face tokenTrouble with Hugging Face trainerTrouble with Hugging Face gemmaTrouble with Hugging Face flightHugging Face integrations break or are missing
History
Trouble with Hugging Face size (3 posts)