← Back to Explorer
Pending

Vicuna-7b-v1.5 (2023)

Vicuna-7b-v1.5 is a large language model released by LMSYS, released under the LLaMA 2 community license license, with a claimed knowledge cutoff of 2023-08-01. Its weights, configuration, and tokenizer files have been cryptographically hashed and are queued for on-chain attestation by Weight Registry.

Description

LMSYS's instruction-tuned LLaMA fine-tune using shared ChatGPT conversation logs, an early demonstration of cheap fine-tuning approaching much larger closed models.

Metadata

Type
LLM
Issuer
LMSYS
License
LLaMA 2 community license
Size
12.55 GB
Knowledge cutoff
2023-08-01
Source
https://huggingface.co/lmsys/vicuna-7b-v1.5
Revision
3321f76e3f527bd14065daf69dad9344000a201d

Hugging Face stats

63,027 downloads on Hugging Face · 403 likes · pipeline: text-generation · library: transformers

transformerspytorchllamatext-generationtext-generation-inference

Weight Hashes

Weights file
2 shards (Merkle root) · 12.55 GB
SHA-256c32945613893504c0ace5443a36c604b9786e2c7467c98c6328c745c283d76d5
SHA3-25628d73eb29a5cb8f87f15ee4d1e77baa2bc2b938ae43acef15a088b42e1e7246d
BLAKE39ff7005249f32a3b332c771bc5501bc9f086002489a4915caba3e7bc5c39ef9a
Config file
config.json · 615 B
SHA-256e122b598d0734b489590e99e3b3562a11ce67ea13bce390a7868b4f73ae6e615
SHA3-25607fa15047eec19b4053e89d11a4936e6d7248e3af5f0f469d1879ffc18ff2aaa
BLAKE3f4d461f765c486398c0cb7cbee314928be04b7ecd83e7d515167a1cb5c1836a6
Tokenizer file
tokenizer.model · 488.01 KB
SHA-2569e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
SHA3-25613d7e7b5b6421b4337cab460f3d9ae8efc9e39faac43bfc184e51cfa01e12c65
BLAKE3e28cb0633a96142c4fb61b6eb9ea581fc066a2e2dd6ea0f98c05a8cd4642ca5f

Verify it yourself

Download the file(s) from the source below, then hash them locally (sha256sum <file> on Linux/macOS, certutil -hashfile <file> SHA256 on Windows) and compare the output against the hash(es) shown above.

Anomalies Flagged

  • sharding signals disagree: index file present=False, filename pattern matched=True -- picked whichever signal is available, but this repo's layout is unusual