Meta Research bitsandbytes Issues

facebookresearch/bitsandbytes Issues

Support for Tesla Architecture
Closed 3 years ago8
OSError: We couldn't connect to 'https://huggingface.co' to load this file
Updated a year ago
is it work on Win platform?
Updated a year ago17
Can I quantize the tensors to other data type rather than 8bit tensor?
Updated a year ago
Quantization functions test fail on Pascal
Updated a year ago1
python setup.py install error
Closed 3 years ago11
Llama?
Updated a year ago
'NoneType' object has no attribute 'cdequantize_blockwise_cpu_fp32'
Updated 2 years ago
Unable to install via Pypi and source
Closed 2 years ago10
8-bit optimizer crashes when fine-tuning gpt2-large
Updated 2 years ago
bfloat16 grads are not supported
Updated 2 years ago
Did you ever try MNMT systems?
Closed 2 years ago9
no difference in memory usage
Closed 2 years ago3
import bitsandbytes as bnb 错误
Closed 2 years ago4
Is AdamW8bit compatible with OSS in fairscale?
Closed 2 years ago1
undefined symbol: __fatbinwrap_38_cuda_device_runtime_compute_75_cpp1_ii_8b1a5d37
Closed 3 years ago6
[Question] Usage of bnb.nn.Embedding with existing classes from other libraries
Closed 3 years ago2
bnb.optim.AdamW
Closed 3 years ago2
How to use it in huggingface AdamW optimizer?
Closed 3 years ago4
Optimizer2State: Unsafe use of eval() in __init__
Closed 3 years ago1
sparse is not passed in __init__ function of StableEmbedding
Closed 3 years ago1
errors when training to the third epoch. everytime.
Closed 3 years ago2
The code uses more GPU memory with Multi-scale Vision Transformers
Closed 3 years ago2
Ok
Closed 3 years ago
Feature request: Please, add implementation for Novograd algorithm
Updated 3 years ago1
undefined symbol: __fatbinwrap_38
Closed 3 years ago2

facebookresearch / bitsandbytes

facebookresearch/bitsandbytes Issues

Support for Tesla Architecture

OSError: We couldn't connect to 'https://huggingface.co' to load this file

is it work on Win platform?

Can I quantize the tensors to other data type rather than 8bit tensor?

Quantization functions test fail on Pascal

python setup.py install error

Llama?

'NoneType' object has no attribute 'cdequantize_blockwise_cpu_fp32'

Unable to install via Pypi and source

8-bit optimizer crashes when fine-tuning gpt2-large

bfloat16 grads are not supported

Did you ever try MNMT systems?

no difference in memory usage

import bitsandbytes as bnb 错误

Is AdamW8bit compatible with OSS in fairscale?

undefined symbol: __fatbinwrap_38_cuda_device_runtime_compute_75_cpp1_ii_8b1a5d37

[Question] Usage of bnb.nn.Embedding with existing classes from other libraries

bnb.optim.AdamW

How to use it in huggingface AdamW optimizer?

Optimizer2State: Unsafe use of eval() in init

sparse is not passed in init function of StableEmbedding

errors when training to the third epoch. everytime.

The code uses more GPU memory with Multi-scale Vision Transformers

Ok

Feature request: Please, add implementation for Novograd algorithm

undefined symbol: __fatbinwrap_38