8 Commits (3d8d5ee5d54102dd73856fac3a80922ea3104a06)

Author SHA1 Message Date
hiyouga dca27b4412 add logits processor 2 years ago
hiyouga ed6161fa6a remove unused code 2 years ago
hiyouga e3aaef7d4a fix layer norm name in PPO 2 years ago
hiyouga 50d9a20f81 alter rewards data type 2 years ago
hiyouga 740a5daf56 support BLOOM models 2 years ago
hiyouga 166c837b95 tiny fix 2 years ago
hiyouga 0c9fda01e3 use fp16 model, add logcallback 2 years ago
hiyouga 769c6ab56b Initial commit 2 years ago