From b246f9bed75b7f96d5be8cf3f6328b0912136f0e Mon Sep 17 00:00:00 2001 From: Kaihui-intel Date: Tue, 4 Mar 2025 15:04:09 +0800 Subject: [PATCH 1/4] add gptq g_idx faq Signed-off-by: Kaihui-intel --- docs/source/faq.md | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/docs/source/faq.md b/docs/source/faq.md index 949b8ce21fc..cc8f97d9c58 100644 --- a/docs/source/faq.md +++ b/docs/source/faq.md @@ -26,9 +26,18 @@ Intel TBB is required, try: $ conda/pip install tbb ``` **Solution:** It's actually already installed by `requirements_pt.txt`, so just need to set up with `export LD_LIBRARY_PATH=/usr/local/lib/:$LD_LIBRARY_PATH`. -#### Issue 6: +#### Issue 7: torch._C._LinAlgError: linalg.cholesky: The factorization could not be completed because the input is not positive-definite. **Solution:** This is a known issue. For more details, refer to [AutoGPTQ/AutoGPTQ#196](https://github.com/AutoGPTQ/AutoGPTQ/issues/196). Try increasing `percdamp` (percent of the average Hessian diagonal to use for dampening), or increasing `nsamples` (the number of calibration samples). +#### Issue 6: +If you run GPTQ quantization with transformers-like API on xpu device, then you may encounter the following error: +```shell +[ERROR][modeling_auto.py:128] index 133 is out of bounds for dimension 0 with size 128 +[ERROR][modeling_auto.py:129] Saved low bit model loading failed, please check your model. +HINT: +XPU device does not support `g_idx` for GPTQ quantization now. Please stay tuned. +You can set desc_act=False. +``` \ No newline at end of file From 33eaff4e2615105ed743e7e570b5968eab5092a8 Mon Sep 17 00:00:00 2001 From: Kaihui-intel Date: Tue, 4 Mar 2025 15:05:48 +0800 Subject: [PATCH 2/4] fix number Signed-off-by: Kaihui-intel --- docs/source/faq.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/source/faq.md b/docs/source/faq.md index cc8f97d9c58..85e0020088a 100644 --- a/docs/source/faq.md +++ b/docs/source/faq.md @@ -26,13 +26,13 @@ Intel TBB is required, try: $ conda/pip install tbb ``` **Solution:** It's actually already installed by `requirements_pt.txt`, so just need to set up with `export LD_LIBRARY_PATH=/usr/local/lib/:$LD_LIBRARY_PATH`. -#### Issue 7: +#### Issue 6: torch._C._LinAlgError: linalg.cholesky: The factorization could not be completed because the input is not positive-definite. **Solution:** This is a known issue. For more details, refer to [AutoGPTQ/AutoGPTQ#196](https://github.com/AutoGPTQ/AutoGPTQ/issues/196). Try increasing `percdamp` (percent of the average Hessian diagonal to use for dampening), or increasing `nsamples` (the number of calibration samples). -#### Issue 6: +#### Issue 7: If you run GPTQ quantization with transformers-like API on xpu device, then you may encounter the following error: ```shell [ERROR][modeling_auto.py:128] index 133 is out of bounds for dimension 0 with size 128 From 4cecc1832162142446c2af5ef78379c1e902d6ba Mon Sep 17 00:00:00 2001 From: "pre-commit-ci[bot]" <66853113+pre-commit-ci[bot]@users.noreply.github.com> Date: Tue, 4 Mar 2025 08:04:36 +0000 Subject: [PATCH 3/4] [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --- docs/source/faq.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/source/faq.md b/docs/source/faq.md index 85e0020088a..d19ec4040e1 100644 --- a/docs/source/faq.md +++ b/docs/source/faq.md @@ -40,4 +40,4 @@ If you run GPTQ quantization with transformers-like API on xpu device, then you HINT: XPU device does not support `g_idx` for GPTQ quantization now. Please stay tuned. You can set desc_act=False. -``` \ No newline at end of file +``` From c9019756885b31a58ffcf8351af2194ca8f35ac0 Mon Sep 17 00:00:00 2001 From: "Sun, Xuehao" Date: Tue, 4 Mar 2025 16:15:02 +0800 Subject: [PATCH 4/4] Add UnicodeEncodeError FAQ Signed-off-by: Sun, Xuehao --- docs/source/faq.md | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/docs/source/faq.md b/docs/source/faq.md index d19ec4040e1..cc38ab1ea81 100644 --- a/docs/source/faq.md +++ b/docs/source/faq.md @@ -41,3 +41,10 @@ HINT: XPU device does not support `g_idx` for GPTQ quantization now. Please stay tuned. You can set desc_act=False. ``` +#### Issue 8: +UnicodeEncodeError: 'charmap' codec can't encode character '\u2191' in position 195: character maps to +**Solution:** +``` +set PYTHONIOENCODING=UTF-8 # for windows +export PYTHONIOENCODING=UTF-8 # for linux +```