rmihaylov falcontune issues

rmihaylov / falcontune

Tune any FALCON in 4-bit

Apache License 2.0

468 stars 51 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Update README.md

#38 Andrew-Cain opened 1 year ago
0
FalconLite support

#37 hvico opened 1 year ago
1
Optimizing Inference Time for Chat Conversations on Falcon

#36 HumzaSami00 opened 1 year ago
0
Inference not accurate

#35 Hamad-exarta opened 1 year ago
0
How can I fine-tune with plain sentences instead of pairs of instruction/output or question/answer?

#34 lucas0 opened 1 year ago
0
8-bit models

#33 zepmck opened 1 year ago
0
Its working very well but 2 issues:

#32 TeaCult opened 1 year ago
0
Runtime Error : CUDA OUT OF MEMORY

#31 amnasher opened 1 year ago
0
RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0!

#30 yufengzhe1 opened 1 year ago
6
Is there any way to load the fine-tuned model to huggingface pipeline?

#29 lynngao opened 1 year ago
0
Does pretrained falcon-40b works on colab?

#28 imthebilliejoe opened 1 year ago
0
RuntimeError: expected scalar type Half but found Char

#27 cmazzoni87 opened 1 year ago
0
Added input argument for Alpaca generation

#26 ThiagoViek closed 1 year ago
1
Add contextual generate

#25 koonweee closed 1 year ago
0
ValueError: FP16 Mixed precision training with AMP or APEX (`--fp16`) and FP16 half precision evaluation (`--fp16_full_eval`) can only be used on CUDA devices

#24 chintan-donda opened 1 year ago
1
Metal support

#23 tmc opened 1 year ago
0
RuntimeError: CUDA error: an illegal memory access was encountered

#22 gpravi opened 1 year ago
2
ModuleNotFoundError: No module named 'torch._six'

#21 gpravi opened 1 year ago
0
Does fine-tuning support the multi-GPU training?

#20 cahuja1992 opened 1 year ago
5
OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB (GPU 0; 79.35 GiB total capacity; 77.18 GiB already allocated; 57.19 MiB free; 77.97 GiB reserved in total by PyTorch)

#19 gpravi opened 1 year ago
4
`batch_size, seq_length = input_ids.shape ValueError: not enough values to unpack (expected 2, got 1)`

#18 MohamedAliRashad opened 1 year ago
1
what's the lora target_modules names?

#17 lucasjinreal opened 1 year ago
0
finetune on 4096 context length

#16 MohamedAliRashad opened 1 year ago
2
Support multiturn chat ?

#15 631068264 opened 1 year ago
0
Use prompt get error

#14 631068264 opened 1 year ago
0
You are using a model of type RefinedWeb to instantiate a model of type RefinedWebModel. This is not supported for all configurations of models and can yield errors.

#13 yuhai-china opened 1 year ago
1
generate get error: RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0!

#12 631068264 opened 1 year ago
4
RuntimeError: No available kernel. Aborting execution.

#11 RealCalumPlays opened 1 year ago
7
Could make it being compatible with QloRA ?

#10 znsoftm opened 1 year ago
0
Can give a example to merge the model and lora?

#9 631068264 opened 1 year ago
0
getting error `OSError: libcurand.so.10: cannot open shared object file: No such file or directory` when finetuning

#8 sambar1729 closed 1 year ago
1
TypeError: Input tensors need to be on the same GPU, but found the following tensor and device combinations

#7 631068264 opened 1 year ago
1
Missing compatibility with with torch 1.13

#6 phisad opened 1 year ago
0
bitsandbytes/libbitsandbytes_cpu.so: undefined symbol: cget_col_row_stats

#5 631068264 closed 1 year ago
1
multi-gpu lora finetune falcon

#4 acadaiaca opened 1 year ago
10
Inference speed for 7B models (triton backend, GTX 3090)

#3 nikshepsvn opened 1 year ago
3
Error when trying to run 7B model

#2 nikshepsvn closed 1 year ago
1
Possible to offload to cpu (ram)

#1 angelovAlex opened 1 year ago
5