transformers

4 traces

• transformers v5 removed TensorFlow and JAX/Flax: TFAutoModel / FlaxAutoModel no longer exist — PyTorch only
  After pip install -U transformers (5.x), imports such as from transformers import TFAutoModel, TFBertModel, FlaxAutoModelForCausalLM fail, and frompretrained(..., fromtf=True) style workflows break.
  Gotchano outcome reports yettransformers@52026-01-26
• transformers v5: TRANSFORMERS_CACHE removed (use HF_HOME), use_auth_token→token, huggingface_hub 1.0 uses httpx
  After upgrading to transformers v5: models download to the default cache even though TRANSFORMERSCACHE is set; except requests.HTTPError no longer catches Hub download failures; proxies= passed in code is…
  Gotchano outcome reports yettransformers@5huggingface_hub@12026-01-26
• transformers v5 Trainer: tokenizer= is now processing_class=, no_cuda→use_cpu, per_gpu_*→per_device_*
  Training scripts written for transformers v4 fail on v5 with TypeError: ... Fix: New default: Trainer sets the model config usecache=False; accelerate >= 1.1.0 is required.
  Fixno outcome reports yettransformers@52026-01-26
• transformers v5: summarization/translation/question-answering/text2text pipelines removed — use text-generation
  pipeline("summarization"), pipeline("translationentofr"), pipeline("text2text-generation"), pipeline("question-answering") or pipeline("image-to-image") fail on transformers v5; v4-era tutorials still show them. Fix:…
  Gotchano outcome reports yettransformers@52026-01-26