GitHub / On-Point-RND/GIFT_SW / commits
| SHA | Message | Author | Date | Stats |
|---|---|---|---|---|
| d2385b31 | Update README.md |
Viktor Moskvoretskii <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
over 1 year ago | |
| e466ac05 | Update README.md |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 408bee6d | Update README.md |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2af9cf77 | merge Merge branch 'huggingface:main' into main |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 850eeb5c | FIX Pre-commit version in config (#2034) |
ltoniazzi <6****i@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 5996d394 | TST Enable more tests in XPU (#2036) |
Fanli Lin <f****n@i****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 900f96c4 | [Add] DoRA Embedding (#2006) |
Aritra Roy Gosthipaty <a****y@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| c3b63ce2 | ENH Test and DoRA compatibility with XPU |
Fanli Lin <f****n@i****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 1a5d0f81 | FIX: Don't target the classification head when using target_modules="all-line... |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| d3023f3f | Update README.md |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| fa564f28 | Update README.md |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| a95d3124 | merge Merge branch 'huggingface:main' into main |
Vitya Vitalich <v****y@y****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 70eaccd4 | add GIFT-SW adapter | VityaVitalich <v****y@y****u> | about 2 years ago | |
| f3c7c6e5 | ENH Raise error when applying modules_to_save on tuner layer (#2028) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 8fcb1951 | MAINT: Update ruff version to ~0.6.1 (#1965) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| fa218e19 | TST test_mixed_adapter_batches_lora_opt_timing on XPU (#2021) |
Fanli Lin <f****n@i****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 6c832c1d | TST Make TestModelAndLayerStatus device-agnostic (#2026) |
Fanli Lin <f****n@i****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 95821e5c | ENH: Better error msg for replace_lora_weights_loftq when using a local model... |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 25ab6c9b | TST Enable regression tests on XPU (#2019) |
Fanli Lin <f****6@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| b4cf1b3c | CI Remove regression tests from BNB CI (#2024) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| eb5eb6ef | TST Enable test_vera_dtypes on XPU with bf16 (#2017) |
Fanli Lin <f****6@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| f71e89f7 | FIX Deprecated params/funcs in X-LoRA example (#2010) |
Eric Buehler <6****r@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| e8ba7de5 | CI Activate single core multi backend bnb tests (#2008) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 0222450f | TST: Potentially Skip 8bit bnb regression test if compute capability is too l... |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 4c3a76fa | FIX DOC Update X-LoRA docs, some bugfixes (#2002) |
Eric Buehler <6****r@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 670d0fac | FIX CI Correctly report outcome of bnb import test (#2007) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 22f042a1 | ENH: Warn when a user provided model name in the config renamed (#2004) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| d6e772f1 | TST Add LNTuningConfig and LoKrConfig to tests (#2005) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 04212346 | DOC Fix typos in lora.md (#2003) |
nir-sh-automat-it <1****t@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 41c274ec | FIX Import error in BOFT half precision test (#1995) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 9988cb9d | FIX BOFT, OFT saving merged layers (#1994) |
Zeju1997 <4****7@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| fcac30be | MAINT Default to loading weights_only for torch (#1993) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2a5d3132 | ENH Small updates to helper.rescale_adapter_scale (#1989) |
Aritra Roy Gosthipaty <a****y@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| c8696648 | FIX BOFT mixed precision (#1925) |
Edenzzzz <w****n@w****u>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 4611034f | FIX: Adjust transformers version check for bloom (#1992) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| b9260305 |
FIX Docker build CI (#1987)
Signed-off-by: Adrien <a****n@h****o> |
Adrien <a****n@h****o>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| f5142831 | DOC Docs and examples for X-LoRA (#1970) |
Eric Buehler <6****r@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 9a087823 | DOC Small fixes for HQQ and section title (#1986) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 46f78978 | FEAT Context manager for scaling LoRA (#1951) |
Aritra Roy Gosthipaty <a****y@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 269aba53 | ENH AdaLoRA: warn when user use r argument (#1981) |
Bhargav Yagnik <4****k@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 52a4ac9c | ENH Faster bf16 merging on CPU (#1978) |
Saaketh Narayan <n****h@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| c874ba3f |
CHORE Update CI configuration for workflows (#1985)
Signed-off-by: Adrien <a****n@h****o> |
Adrien <a****0@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| f13d860e | FIX Loading adapter honors offline mode (#1976) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| f6d3e386 | FIX active_adapters for transformers models (#1975) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 7e7b5588 | FIX: lora+: include lr in optimizer kwargs (#1973) |
kallewoof <k****m@g****p>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 1b16753a | ENH Update VeRA preconfigured models (#1941) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 27833a2e | FIX: New bloom changes breaking prompt learning (#1969) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 273acf05 |
FEAT: Add LoRA+ (#1915)
Co-authored-by: moghadas76 <s****2@g****m>, Chris Hua <s****c@u****m> |
kallewoof <k****m@g****p>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 296fbcde | FIX Prefix tuning if past_key_values is passed (#1942) |
Nidhogg-lyz <7****z@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| f2b6d13f | CI Fix Windows permission error on merge test (#1952) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 8aacb993 | Bump version to 0.12.1.dev0 (#1950) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| e6cd24c9 | Release v0.12.0 (#1946) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 05f57e94 | PiSSA, OLoRA: Delete initial adapter after conversion instead of the active a... |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2ce83e05 | FIX Decrease memory overhead of merging (#1944) |
Saaketh Narayan <s****h@m****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| ebcd0792 |
[WIP] ENH Add support for Qwen2 (#1906)
Co-authored-by: BenjaminBossan <b****n@g****m> |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| ba75bb14 |
FIX: More VeRA tests, fix tests, more checks (#1900)
Co-authored-by: BenjaminBossan <b****n@g****m> |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 6472061a | FIX Prefix tuning Grouped-Query Attention (#1901) |
Tianwen Tang <4****8@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| e02b938e | FIX PiSSA & OLoRA with rank/alpha pattern, rslora (#1930) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 52684952 | FEAT Add HRA: Householder Reflection Adaptation (#1864) |
DaShenZi <3****1@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2aaf9ced | ENH Sync LoRA tp_layer methods with vanilla LoRA (#1919) |
zhangshengdong29 <4****3@q****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| a019f869 | FIX sft script print_trainable_parameters attr lookup (#1928) |
anch0vy <a****a@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2a6402f4 | DOC Fix typo of encoder_reparameterization_type (#1926) |
Soyoger <8****6@q****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| e72a96f7 |
FEAT Add FourierFT Support (#1838)
Co-authored-by: zqgao22 <z****t@c****k>, Chaos96 <w****7@m****n>, DSAILatHKUST <d****t@1****m> |
Aochuan Chen <p****n@f****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 48e136d9 | FIX: Flaky multitask prompt tuning test fixed by setting the seed (#1908) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 58afb34e | FEAT Integrate X-LoRA (#1491) |
Eric Buehler <6****r@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 01f1b992 | Example: DNA Language Model. (#1873) |
Rahul Shrestha <r****1@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 09358aad | Chore: Docs markdown formatting (#1899) |
Chris Hua <h****r@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 31c0d857 | FIX DeepSpeed recursion error (#1892) |
清晨的雾 <4****1@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 018a1f49 | FIX TEST Even higher tolerance for AdaLoRA in test (#1898) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 1e2258d7 | ENH Ephemeral GPU offload support for DoRA (#1857) |
kallewoof <k****m@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 1e5227ff | TST Bump absolute tolerance for test (#1891) |
kallewoof <k****m@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 62122b5a | FIX TEST Higher tolerance for AdaLoRA in test (#1897) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 9dc53b8f | CI Don't fail fast in test matrix (#1896) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| db8b76fd | DOC DoRA example script & notebook (#1885) |
Shirin Yamani <7****i@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 7ffa43b1 | FIX Avoid early import of torch extension by BOFT (#1879) |
Ting PAN <t****n@s****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 27bc3054 | FIX sft script: only print trainable params if peft (#1888) |
Ma, Guokai <g****a@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 184beaf1 | FIX Make special LoRA inits DeepSpeed compatible (#1887) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| c9b19bb8 | FIX Init AdaLoRA to be identity transform (#1884) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| ef23712b |
ENH: LoRA support for dynamically dispatching to custom layers (#1875)
Co-authored-by: Steven Liu <5****u@u****m> |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| d716adf3 | Update bug-report.yml (#1882) |
Younes Belkada <4****a@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| d37dde61 | FIX Error when using VeRA with fp16 or bf16 (#1874) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 53643514 | CI Downgrade numpy to <2.0 for Mac and Windows (#1871) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 717db6e1 | CI testing BNB: remove single GPU tests (#1866) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 5194aef5 | Attempt to fix the red messages (#1868) |
Younes Belkada <4****a@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 25c0fe9a | FIX fix multitask prompt tuning paper link (#1862) |
Sepehr Rafiei <s****0@g****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| e0e8204b | Update lora_based_methods.md (#1861) |
smurfyknuckles <j****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 076561bb | CI Testing: Remove bnb import check (#1859) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| efda766f | DOC Move helpers section to dev developer guide (#1856) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| d608f832 | DOC FIX Comment about init of LoRA Embedding (#1855) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 19461353 | Update nightly-bnb.yml (#1854) |
Younes Belkada <4****a@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 3831e06a | FIX: Adalora ranknum loaded on wrong device (#1852) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 2f5360a7 | FEAT Add OLoRA initialization strategy to LoRA (#1828) |
Kerim <9****e@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 8843a767 | MNT Upgrade ruff version to ~0.4.8 (#1851) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| b6af7feb | DOC Fix PeftMixedModel docstring example #1824 (#1850) |
Naman Vats <3****s@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 47b3d742 | CI Activate env to prevent bnb import error (#1845) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 7b1c08d2 | ENH Support different layer shapes for VeRA (#1817) |
Dawid <2****i@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| a8286a7b | DOC Describe torch_device in from_pretrained docs (#1843) |
Benjamin Bossan <B****n@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 683db0fa | feat(ci): add trufflehog secrets detection (#1841) |
Luc Georges <M****e@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 0f89d34d | Fix broken messages (#1842) |
Younes Belkada <4****a@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago | |
| 0b40d1a3 | Workflow / Bnb: Add a mechanism to inform us if the import fails (#1830) |
Younes Belkada <4****a@u****m>
Committed by: GitHub <n****y@g****m> |
about 2 years ago |