dolly

dolly copied to clipboard

OOM issue when finetune with V100

11

I tried to run the training.trainer script with batch size == 1 (originally it is 8), but met OOM issue with V100. Has anyone tried to finetune it with V100-32G?...

bingjie3216

ValueError: Your setup doesn't support bf16/gpu.

1

I run the notebook on 8 V100 GPUs, but an error occured: ``` File "", line 105, in __init__ File "/databricks/python/lib/python3.9/site-packages/transformers/training_args.py", line 1098, in __post_init__ raise ValueError( ValueError: Your setup...

yinwangsong

Running the code without databricks

6

Hi in the `train_dolly.py`, a lot of MAGIC commands were there which is used in databricks notebooks. Do we need to run those commands separately if we are not using...

ayubih

RuntimeError: Could not find response key token IDs when using bloom model and tokenizer to train

3

it work fine if i use gpt-j i guess this because of tokenizer and this https://github.com/databrickslabs/dolly/blob/03bf3852daa42e6091a39483dda0714c02de7382/training/trainer.py#L52 any tips to adjust it so it can use other model than gpt-j ?...

acul3

Not able to select an on Azure any GPU machine

3

Hi. Trying to run dolly on MS Azure. When I try to create compute cluster, and choosing Runtime 12.2 LTS, cannot choose any GPU machine, like Standard_ND96asr_v4. `Error: This node...

jcsjacekj

Please make weights / checkpoint available

1

Otherwise we are all doing the same training each time, its wasteful. Thanks!

pathway

Couldn't find the train_dolly notebook.

2

> Open the `train_dolly` notebook in the `dolly repo`, attach to your GPU cluster, and run all cells. When training finishes, the notebook will save the model under `/dbfs/dolly_training`. In...

ayubih

Using Bigscience Bloom 176B or Bloomz 176B instead of GPT-J 6B

1

Would it be possible to take this software and substitute the Bigscience Bloom 176B or Bloomz 176B models, instead of the present GPT-J 6B model, as a simple drop-in in...

sblaszak

for 12.2 LTS all gpu optimized nodes are disabled

1

I am not able to select p4 as mentioned on github page and standard doesnt even show up. I am using the premium account

allthingssecurity

Hosting the Dolly dataset on the Hugging Face Hub

2

Hi Databricks team, this is a really cool project and great job creating a high quality instruction dataset with a permissive license! Would you be interested in hosting the Dolly...

lewtun