XLNet-base-cased is a large language model released by Google/CMU, released under the MIT license, with a claimed knowledge cutoff of 2019-06-19. Its weights, configuration, and tokenizer files have been cryptographically hashed and are queued for on-chain attestation by Weight Registry.
Description
Permutation-based autoregressive pretraining that factorizes the joint probability over random permutations of the input sequence, capturing bidirectional context without BERT's artificial [MASK] token. Outperformed BERT on 20 NLP tasks at publication, including SQuAD, GLUE, and RACE.
Metadata
Hugging Face stats
217,319 downloads on Hugging Face · 82 likes · pipeline: text-generation · library: transformers
Weight Hashes
Weights file
SHA-256b13dc2d3664a385b92d087229d5410d9e3020ede976e7ff4c62c9c85cd969a42
SHA3-256b157bcf86f3cc750c823af5ee0286e42d002354b83afa4a2eeb47b129c61b349
BLAKE3732dba6cc986d3c27d3ceaeffbd26bfc8b74b6621e112c226961a04c5ed7b60f
Config file
SHA-2560fde1d9aa1bfabdd3af475ef5811a093ead6b13f443becc06f426a8dba214531
SHA3-2563ba6a4ff29d07baec9f5476e91165d698261fa097b995af5c0b93d0564474659
BLAKE389b449d3e73902ca0de2164233a3c0f7844fbbc21dd72fcea897bb2448002492
Tokenizer file
SHA-25653fa8c1e443eda9cfc88c906d9181256e582638dbd5ac3ba00d73901ef04f107
SHA3-2568453a8bcb79d971e2bb0a333ef4dd8e9ea521dd33bb391a6d2ad80624a1b6a38
BLAKE3f2ae335b9590851b1a4779df38b94b873682178ce63cc732250ea27d4fd420cf
Verify it yourself
curl -L -o "pytorch_model.bin" "https://huggingface.co/xlnet/xlnet-base-cased/resolve/ceaa69c7bc5e512b5007106a7ccbb7daf24b2c79/pytorch_model.bin" sha256sum "pytorch_model.bin"
Expected SHA-256: b13dc2d3664a385b92d087229d5410d9e3020ede976e7ff4c62c9c85cd969a42