Help evaluating the model

#29
by AIC-MCIT - opened

we really want to know the results of our submitted model, so can we know the process of the evaluation and when can we expect to see the results

Open Arabic LLM Leaderboard org

can we ask when will the evaluation process ends and when will it be published on the leaderboard

Open Arabic LLM Leaderboard org

@AIC-MCIT Thanks for submitting the model to the leaderboard
For one reason or another, your model seems to be slow with our eval pipeline at least with Alrage (generative task), are you experiencing something similar? will update you here Im able to fix it!
btw other benchmarks (Apart from alrage) are actually done and can be seen in here

the model is MOE so inferring it with transformers only could be slow as it needs some tweaks to handle MOE , so i recommend using VLLM for generating the responses it will be much faster

Open Arabic LLM Leaderboard org

noted @AIC-MCIT
your model now is in the leaderboard! (congrats in being TOP)
sorry for the delay

amztheory changed discussion status to closed

Sign up or log in to comment