huggingface safetensors issues

huggingface / safetensors

Simple, safe way to store and distribute tensors

https://huggingface.co/docs/safetensors

Apache License 2.0

2.79k stars 189 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Can't save checkpoint to file descriptor.

#531 GLivshits opened 15 hours ago
0
fix(save_file): add force_contiguous(default=True) to save_file

#530 dbyoung18 opened 1 week ago
1
let `safetensors.load` support load weight from bytearray

#529 JeryZeng opened 1 week ago
0
RuntimeError: unable to mmap 23802932552 bytes from file <./flux1-dev.safetensors>: Cannot allocate memory (12)

#528 sakura-nyaa opened 1 week ago
0
[Question] Comparison with the zarr format?

#527 julioasotodv opened 1 week ago
0
Improve clarity in safetensors' `index.mdx` doc

#526 ethanknights opened 3 weeks ago
0
Prepare for 0.4.5

#525 Narsil closed 3 weeks ago
0
add remapping to safetensors.torch.load_model

#524 f-fuchs opened 1 month ago
0
SafetensorError does not appear in `__init__.pyi`

#523 alanhdu opened 1 month ago
1
Retrieving saved metadata

#521 parlance-zz closed 1 month ago
3
Some layers are not loaded(‘encoder.embed_token.weight’ and 'shared.weight') when loading converted safetensors into transfomrers.MT5EncoderModel.

#520 WongGawa opened 1 month ago
0
converting weight of more than 50G‘s .bin to .safetensors failed

#519 WongGawa opened 1 month ago
0
fix lifetimes, add tensors iterator

#518 gvilums closed 3 weeks ago
3
Use id rather than modelId

#517 osanseviero closed 1 month ago
0
Adding support for wrapper tensor subclass for safetensor

#516 jerryzh168 closed 2 weeks ago
6
Problem serializing quantized weights

#515 sayakpaul closed 1 month ago
3
Forcing empty list/empty tuple behavior.

#514 Narsil closed 1 month ago
0
Fixing release workflow.

#513 Narsil closed 1 month ago
0
Git v0.4.4 rc0

#512 Narsil closed 1 month ago
0
Upgrading the version.

#511 Narsil closed 1 month ago
0
Fixing zero-width tensor for in memory loading.

#510 Narsil closed 2 months ago
1
Respects torch.device(0) new behavior without breaking backward compatibility

#509 Narsil closed 2 months ago
2
Change doc-builder workflow container

#508 mishig25 closed 2 months ago
2
Improving bf16 tests

#507 Narsil closed 2 months ago
2
Revert "Improving the bf16 tests for PT+TF."

#506 Narsil closed 2 months ago
0
Improving the bf16 tests for PT+TF.

#505 Narsil closed 2 months ago
0
Upgrading Pyo3 0.22 (with gil-refs)

#504 Narsil closed 2 months ago
0
Wrong device when using device="cpu" with torch.device

#503 chrisdesa opened 2 months ago
3
Pr

#502 goelayu closed 2 months ago
0
Add `safetensors.torch.save_state_dict()` to python bindings

#501 mohahf19 closed 2 months ago
1
Query device name from pytorch if only device index is given

#500 dvrogozh closed 2 months ago
11
pytorch: safetensors library hardcodes using CUDA if only device index is provided

#499 dvrogozh closed 2 months ago
3
get_tensor returns an empty tensor on LoRa safetensors adapter saved by SFTTrainer

#498 tsvisab closed 2 months ago
1
Add conversion validation support for the float8 tensor dtype

#497 darthhexx closed 2 months ago
1
get_slice has the same speed as get_tensor

#496 youkaichao closed 2 months ago
1
inconsistent behavior of slicing in certain model file

#495 youkaichao closed 1 month ago
4
Add support for xla devices

#494 BitPhinix closed 2 months ago
1
SafetensorError is not pickleable, potentially causing hangs in multi-process environments

#492 Wehrdo closed 2 months ago
2
Compilation fail due to test error

#491 BananPrzydawka closed 2 months ago
3
How to save model checkpoint from a distributed training from multiple nodes?

#490 Emerald01 closed 2 months ago
2
Unable to save llama3 since complex64 is not supported

#489 npuichigo closed 2 months ago
1
Minor correction about Arrow in README.md

#488 lhoestq closed 2 months ago
1
README is inaccurate

#487 judahrand closed 2 months ago
3
Register a MIME type for the `safetensors` format.

#486 judahrand closed 2 months ago
1
misleading error message in `safe_open`

#485 Clement-Lelievre closed 2 months ago
4
feat(ci): add trufflehog secrets detection

#484 McPatate closed 2 months ago
0
ModuleNotFoundError: No module named 'torch._higher_order_ops'

#482 guotong1988 closed 4 months ago
1
Error when using map_location="cuda:0" with multiple GPUs

#481 AAnirudh07 closed 3 months ago
1
safetensors.tensorflow.save modifies its input argument

#480 allaisandrea closed 3 months ago
1
support for Cambricon MLUs device

#479 huismiling closed 3 months ago
1