-
Notifications
You must be signed in to change notification settings - Fork 143
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Mistral rs python binding error #323
Comments
Thank you for reporting this. Can you please try to run an example from the command line, for example the below, which loads a GGUF model as configurable in the .toml file here:
I think that if this works, we can isolate the issue with the Python bindings and not a general issue.
You can run |
Thank you for your response. I tried running: But I get the below error: .\mistralrs_server : The term '.\mistralrs_server' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is
I am guessing I am either not running it from the right directory or I need to add something to PATH before running it. Can you tell me where I must run you |
Ok, sorry I wasn't more clear. If you go to the
|
I was able to run the command. I got an error, this is my logs from running the command : cargo run --release --features cuda -- -i toml -f toml-selectors/gguf.toml Updating crates.io index |
@shresht8, sorry for the wait. coreylowman/cudarc#240 should fix this, I'll let you know when it gets merged. |
Hi @shresht8, I think it should be fixed now. Can you please try it again after a |
Bug:
I am attempting to run mistral rs for inference for my own GGUF files but before that I wanted to test with the example given in the documentation. I installed mistral rs using ubuntu 22.04 since I have a windows machine with RTX 4070 GPU. I installed the dependencies required as well using ubuntu such as the libssl-dev and pkg-config. Once that was done I went into my conda environment and ran the pip install mistralrs-cuda command to install the package. It ran successfully without any errors so I assumed the installation was complete. Is there any verification process for the installation?
Once the installation was complete, I successfully imported the packages and then attempted to run the example given in the documentation given for the python bindings:
runner = Runner(
which=Which.GGUF(
tok_model_id="mistralai/Mistral-7B-Instruct-v0.1",
quantized_model_id="TheBloke/Mistral-7B-Instruct-v0.1-GGUF",
quantized_filename="mistral-7b-instruct-v0.1.Q4_K_M.gguf",
tokenizer_json=None,
repeat_last_n=64,
)
)
and got the below error:
PanicException Traceback (most recent call last)
Cell In[2], line 1
----> 1 runner = Runner(
2 which=Which.GGUF(
3 tok_model_id="mistralai/Mistral-7B-Instruct-v0.1",
4 quantized_model_id="TheBloke/Mistral-7B-Instruct-v0.1-GGUF",
5 quantized_filename="mistral-7b-instruct-v0.1.Q4_K_M.gguf",
6 tokenizer_json=None,
7 repeat_last_n=64,
8 )
9 )
PanicException: called
Result::unwrap()
on anErr
value: LoadLibraryExW { source: Os { code: 126, kind: Uncategorized, message: "The specified module could not be found." } }I attempted to run the actual mistral model as well (not the GGUF version) like this:
runner = Runner(
which=Which.Plain(
model_id="mistralai/Mistral-7B-Instruct-v0.1",
arch=Architecture.Mistral,
tokenizer_json=None,
repeat_last_n=64,
)
)
But I got the same error as above. Can anyone help me understand the cause of this error?
Latest commit
commit 455653c (HEAD -> master, origin/master, origin/HEAD)
The text was updated successfully, but these errors were encountered: