Add TF h5 weights (converted from PT, so TF-only users skip from_pt path)

add vocab.json for from_pretrained_via_hub fallback

add tokenizer_config.json for from_pretrained_via_hub fallback

add tokenizer.json for from_pretrained_via_hub fallback

add tf_model.h5 for from_pretrained_via_hub fallback

add merges.txt for from_pretrained_via_hub fallback

add generation_config.json for from_pretrained_via_hub fallback

add flax_model.msgpack for from_pretrained_via_hub fallback

add config.json for from_pretrained_via_hub fallback

Add attribution README — credit: OpenAI / HuggingFace

Mirror of openai-community/gpt2 SHA-256:7c5d3f4b8b76583b422fcb9189ad6c89d5d97a094541ce8932dce3ecabde1421

initial commit