issues
search
johnsmith0031
/
alpaca_lora_4bit
MIT License
534
stars
84
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Support for moe model？
#156
laoda513
opened
7 months ago
2
Why lora support is only for simple lora with only q_proj and v_proj ？
#155
XpracticeYSKM
opened
8 months ago
1
trying to get this working with text-generation-webui
#154
caterpillarpants
opened
1 year ago
9
AttributeError: 'dict' object has no attribute 'to_dict'
#153
kkaarrss
closed
1 year ago
1
Finetuning CodeLLaMA34B - RuntimeError: The size of tensor a (1024) must match the size of tensor b (8192)
#152
juanps90
closed
1 year ago
3
docker.io/nvidia/cuda:11.7.0-devel-ubuntu22.04 not available anymore
#151
kkaarrss
closed
1 year ago
1
3 errors detected in the compilation of "src/alpaca_lora_4bit/quant_cuda/quant_cuda_kernel.cu"
#150
kkaarrss
closed
1 year ago
2
Refactor generate_prompt to f-strings and make input optional
#149
jordankzf
closed
1 year ago
1
ValueError: Target module Autograd4bitQuantLinear() is not supported.
#148
jordankzf
closed
1 year ago
7
module 'alpaca_lora_4bit.quant_cuda' has no attribute 'vecquant4recons_v2'
#147
kevkid
closed
1 year ago
4
OOM on inference while i can finetune with more tokens
#146
nepeee
closed
1 year ago
2
Merging LoRA after finetune
#145
gameveloster
opened
1 year ago
1
Unable to Build Wheels
#144
VegaStarlake
closed
1 year ago
8
Is alpaca_lora_4bit@winglian-setup_pip missing finetune.py?
#143
tensiondriven
opened
1 year ago
1
Feature request: Stop when loss reaches X
#142
tensiondriven
opened
1 year ago
1
Targeting all layers and biases
#141
grimulkan
closed
1 year ago
2
High perplexity while lower loss after LoRA finetuning (how?)
#140
alex4321
closed
1 year ago
5
Error attempting to finetune llama2-70b
#139
tensiondriven
opened
1 year ago
5
Flash Attention 2
#138
ghost
closed
1 year ago
1
LoRA Output Identical to Base Model
#137
LegendBegins
closed
1 year ago
4
How to use inference.py after finetune.py?
#136
athenawisdoms
closed
1 year ago
2
Checkpoint saving broken with the latest version of huggingface
#135
nepeee
closed
1 year ago
8
Tests
#134
alex4321
closed
1 year ago
1
Use PyPI versions of dependencies
#133
alex4321
closed
1 year ago
1
July
#132
ghost
closed
1 year ago
4
Crashes during finetuning
#131
gameveloster
opened
1 year ago
2
TypeError: object of type 'NoneType' has no len()
#130
athenawisdoms
closed
1 year ago
1
Update docs for > 2048 token models (SuperHOT)?
#129
tensiondriven
opened
1 year ago
12
Use gradient checkpoint only for training mode, not evaluation
#128
alex4321
closed
1 year ago
2
Gibberish results for non-disabled "faster_mode" using "vicuna-7B-GPTQ-4bit-128g" model
#127
alex4321
opened
1 year ago
4
Zero initializer for biases
#126
alex4321
closed
1 year ago
1
Fix NaN or Inf after initializing Vicuna models (due to lack of bias weights)
#125
alex4321
closed
1 year ago
1
Inf or NaN in probabilities. Windows 10, vicuna-7b-gptq-4bit-128g
#124
alex4321
closed
1 year ago
35
Remove import of custom_autotune (it's triton-dependent, triton depen…
#123
alex4321
closed
1 year ago
2
this repo support 2bit finetuning the llama model？ Is there any case to show how to run the scripts?
#122
zlh1992
opened
1 year ago
1
[question] weights in the replaced quantized modules
#121
vince62s
closed
1 year ago
0
how to change into 8 bit
#120
leexinyu1204
opened
1 year ago
1
Problem with inference
#119
leexinyu1204
closed
1 year ago
7
fine tune with 2 GPU
#118
shawei3000
opened
1 year ago
2
how to infer with finetuned model?
#117
balaji-skoruz
opened
1 year ago
4
Implementing Landmark Attention
#116
juanps90
opened
1 year ago
0
Finetuning 2-bit Quantized Models
#115
kuleshov
opened
1 year ago
7
ImportError: cannot import name '_get_submodules' from 'peft.utils'
#114
saber258
closed
1 year ago
10
Differences between QLoRA and this repo
#113
qwopqwop200
opened
1 year ago
3
Code reference request
#112
PanQiWei
opened
1 year ago
1
what is the difference between v1 model and v2 model?
#111
zlh1992
opened
1 year ago
1
Problem loading safetensor file format
#110
ortegaalfredo
opened
1 year ago
3
WIP: Checkpoint merge of LoRA
#109
Ph0rk0z
opened
1 year ago
4
Support datastets with prompt and response like blue moon or many chat formats.
#108
Ph0rk0z
closed
1 year ago
0
Consider using new QLoRA
#107
juanps90
opened
1 year ago
3
Next