NielsRogge/transformers

Pull Requests Commits

Restore fp16 support on xla gpu device (#22300)

ymwangg committed 3 years ago

Verified d35f7296

Time to Say Goodbye, torch 1.7 and 1.8 (#22291)

ydshieh committed 3 years ago

Verified 67c2dbdb

Add translation perf_infer_gpu_one for it (#22296)

davidegazze committed 3 years ago

Verified 86c7931a

fix more doctests (#22292)

ydshieh committed 3 years ago

Verified d0b942d1

More doctests (#22268)

ydshieh committed 3 years ago

Verified 48327c57

Fix error in mixed precision training of `TFCvtModel` (#22267)

gcuder committed 3 years ago

Verified 5a2b77a6

replace_8bit_linear modules_to_not_convert default value fix (#22238)

Andrei Panferov committed 3 years ago

Verified 330d8b99

Update vision docstring bool masked pos (#22237)

amyeroberts committed 3 years ago

Verified c07a02a4

Example of pad_to_multiple_of for padding and truncation guide & docstring update (#22278)

MKhalusova committed 3 years ago

Verified 7bd86505

Move torch.compile() wrapping after DDP/FSDP wrapping to ensure correct graph breaks during training (#22279)

ani300 committed 3 years ago

Verified fb0a38b4

Fix doc links (#22274)

amyeroberts committed 3 years ago

Verified 8ac29fe0

Proper map location for optimizer load (#22273)

sgugger committed 3 years ago

Verified da005253

Rework a bit the LLaMA conversion script (#22236)

sgugger committed 3 years ago

Verified 786092a3

Fix balanced and auto device_map (#22271)

sgugger committed 3 years ago

Verified 43efd7cb

Fix the gradient checkpointing bug of the llama model (#22270)

yqy2001 committed 3 years ago

Verified 89f0fda5

[Trainer] Add optional communication backends for torch.distributed when using GPU (#22247)

heya5 committed 3 years ago

Verified cf0af9a3

Italian translation perf_infer_cpu (#22243)

nickprock committed 3 years ago

Verified c4bf6f38

[Docs] fix typos in some tokenizer docs (#22256)

yesinkim committed 3 years ago

Verified 466144d4

Update training_args.py -- a nightly install is not required anymore for torch.compile (#22266)

pminervini committed 3 years ago

Verified a48310de

[trainer] param count for deepspeed zero3 (#22193)

stas00 committed 3 years ago

Verified 60d51ef5

Fix Unnecessary move of tensors from CPU to GPU in LlamaRotaryEmbedding (#22234)

ma787639046 committed 3 years ago

Verified cf601b90

Revert "Use `dash==2.8.1` for now for daily CI" (#22233)

ydshieh committed 3 years ago

Verified bec07561

Fix natten (#22229)

alihassanijr committed 3 years ago

Verified 3028b20a

fix(docs): fix task guide links in model docs (#22226)

Seb0 committed 3 years ago

Verified 074490b2

Removed .mdx extension in two links (#22230)

MKhalusova committed 3 years ago

Verified 314cdf7c

Add LlamaForSequenceClassification (#22209)

lewtun committed 3 years ago

Verified f2514413

fix AutoTP in deepspeed could not work for bloom (#22196)

sywangyi committed 3 years ago

Verified 675d2a5a

LLaMA house-keeping (#22216)

sgugger committed 3 years ago

Verified 00934026

Depth estimation task guide (#22205)

MKhalusova committed 3 years ago

Verified 42f8f764

Use `dash==2.8.1` for now for daily CI (#22227)

ydshieh committed 3 years ago

Verified 53218671

Newer Older