Sean Owen comments

Results 245 comments of


                                            Sean Owen

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

Yes at the moment that is the substantial difference.

Not able to select an on Azure any GPU machine

You need to choose a GPU runtime - and _unselect_ "Use Photon"

Not able to select an on Azure any GPU machine

Closing as duplicate of #2

for 12.2 LTS all gpu optimized nodes are disabled

You need to select a 12.2 ML _GPU_ runtimes to see GPU instance types. (Also, and this bugs me, you need to unselect "Use Photon" to see them.)

Please make weights / checkpoint available

I believe the issue is that the weights are a function of the training data, which has a non-commercial license and ultimately stems from OpenAI output. That means downstream users...

Correct one inaccurate Q&A data point "How can I contribute to Spark?"

I'm not sure if we're going to edit the dataset unless it contains harmful or sensitive info (shouldn't!) or obvious content problems. In a way we want to capture the...

Correct one inaccurate Q&A data point "How can I contribute to Spark?"

oh yeah fair, but I'll leave it to maintainers to decide policies on typos like that

Correct one inaccurate Q&A data point "How can I contribute to Spark?"

@matthayes is going to collect this feedback and get it into a larger batch update, thank you

ValueError: Could not load model databricks/dolly-v2-12b with any of the following classes: (, ).

Can you try adding `task="text-generation"` to the `pipeline()`? That's worked for a few other people. @matthayes I think this task `instruction-following` thing might be causing this issue.

ValueError: Could not load model databricks/dolly-v2-12b with any of the following classes: (, ).

As a rule of thumb, because this model is stored in 16-bit, you will need 2 x num parameters bytes of mem, plus memory for the data and so on....