sgsdxzy a3b1602391 fix: rope scaling for cohere and qwen (#436) 9 mesiacov pred
..
__init__.py c2aaaefd57 allow out-of-tree model registry 9 mesiacov pred
baichuan.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
bloom.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
chatglm.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
cohere.py a3b1602391 fix: rope scaling for cohere and qwen (#436) 9 mesiacov pred
dbrx.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
decilm.py e31c6f0b45 feat: refactor modeling logic and support more models (#274) 11 mesiacov pred
deepseek.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
falcon.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
gemma.py 0d0c6b313c fix: linear bias of qkv layers in models (#430) 9 mesiacov pred
gpt2.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
gpt_bigcode.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
gpt_j.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
gpt_neox.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
internlm2.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
llama.py 0d0c6b313c fix: linear bias of qkv layers in models (#430) 9 mesiacov pred
llava.py 4d33ce60da feat: Triton flash attention backend for ROCm (#407) 9 mesiacov pred
mixtral.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
mpt.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
olmo.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
opt.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
phi.py 50c2434267 move megatron to a top-level directory 9 mesiacov pred
qwen.py a3b1602391 fix: rope scaling for cohere and qwen (#436) 9 mesiacov pred
qwen2.py a3b1602391 fix: rope scaling for cohere and qwen (#436) 9 mesiacov pred
qwen2moe.py a3b1602391 fix: rope scaling for cohere and qwen (#436) 9 mesiacov pred
stablelm.py 0d0c6b313c fix: linear bias of qkv layers in models (#430) 9 mesiacov pred