 Alexander Borzunov
|
9703358df0
Fix bugs in _choose_num_blocks() added in #346 (#354)
|
2 жил өмнө |
|
Alexander Borzunov
|
1a78638c02
Test that bitsandbytes is not imported when it's not used (#351)
|
2 жил өмнө |
 justheuristic
|
c511990236
Remove unused import os (#352)
|
2 жил өмнө |
|
Alexander Borzunov
|
e12d4c666b
Spam less in server logs (#350)
|
2 жил өмнө |
|
justheuristic
|
010857a834
Estimate adapter memory overhead in choose_num_blocks() (#346)
|
2 жил өмнө |
|
Alexander Borzunov
|
f605f093f7
Support LLaMA repos without "-hf" suffix (#349)
|
2 жил өмнө |
|
Alexander Borzunov
|
90fbaab61e
Fix Docker build by avoiding Python 3.11 (#348)
|
2 жил өмнө |
|
Alexander Borzunov
|
43acfe52a7
Import petals.utils.peft only when needed to avoid unnecessary import of bitsandbytes (#345)
|
2 жил өмнө |
|
Alexander Borzunov
|
294970fe18
Update Colab link
|
2 жил өмнө |
|
Alexander Borzunov
|
515a5120cb
Mention LLaMA in readme (#344)
|
2 жил өмнө |
 Max Ryabinin
|
13f4e3a88a
Fix convergence issues and switch to LLaMA in the SST-2 example (#343)
|
2 жил өмнө |
 Artem Chumachenko
|
b9f0a5467f
Support peft LoRA adapters (#335)
|
2 жил өмнө |
|
Alexander Borzunov
|
dfc6578c8e
Use bitsandbytes 0.40.0.post4 with bias hotfix (#342)
|
2 жил өмнө |
|
Alexander Borzunov
|
b28f5016ea
Delete deprecated petals.cli scripts (#336)
|
2 жил өмнө |
|
Alexander Borzunov
|
fa095f6461
Use 4-bit for llama by default, use bitsandbytes 0.40.0.post3 (#340)
|
2 жил өмнө |
|
Alexander Borzunov
|
158013a671
Implement direct server-to-server communication (#331)
|
2 жил өмнө |
|
Alexander Borzunov
|
4d9c26fe5c
Allow free_disk_space_for() remove arbitrary files from Petals cache (#339)
|
2 жил өмнө |
|
Alexander Borzunov
|
de930918a0
Support loading blocks in 4-bit (QLoRA NF4 format, disabled by default) (#333)
|
2 жил өмнө |
|
Alexander Borzunov
|
66a47c763e
Require pydantic < 2.0 (2.0 is incompatible with hivemind 1.1.8) (#337)
|
2 жил өмнө |
|
Alexander Borzunov
|
10c72acdf4
Fix warmup steps and minor issues in benchmarks (#334)
|
2 жил өмнө |
|
Alexander Borzunov
|
d126ee3053
Add benchmark scripts (#319)
|
2 жил өмнө |
|
Alexander Borzunov
|
fecee8c4dc
Show license links when loading models (#332)
|
2 жил өмнө |
|
Alexander Borzunov
|
47a2b1ee65
Fix llama's lm_head.weight.requires_grad (#330)
|
2 жил өмнө |
|
Alexander Borzunov
|
7a37513f77
Add AutoDistributed{Model, ModelForCausalLM, ModelForSequenceClassification} (#329)
|
2 жил өмнө |
|
Alexander Borzunov
|
cb3f018f9f
Add LLaMA support (#323)
|
2 жил өмнө |
|
Max Ryabinin
|
5c0733711a
Use number of tokens for attn_cache_size (#286)
|
2 жил өмнө |
|
Max Ryabinin
|
c839173e57
Determine block dtype in a unified manner (#325)
|
2 жил өмнө |
|
Max Ryabinin
|
3e7ae5116d
Remove unused imports and attributes (#324)
|
2 жил өмнө |
|
Alexander Borzunov
|
675bacb592
Bump version to 1.1.5 (#312)
|
2 жил өмнө |
|
Alexander Borzunov
|
e026952338
Abort speedtest if it runs too long (#316)
|
2 жил өмнө |
