Closed Flake74 closed 6 months ago
There are missing ( https://github.com/replicate/replicate-go/blob/main/prediction.go ) metrics like input_token_count, output_token_count, total_time for some chat basic models https://replicate.com/meta/llama-2-70b-chat?input=http&output=json
Hi @Flake74. Thanks for calling this out. Before per-token language models, prediction metrics didn't have much information. But now, there's a lot more.
I just opened up #53, which adds this new information.
This is now available in v0.18.0
There are missing ( https://github.com/replicate/replicate-go/blob/main/prediction.go ) metrics like input_token_count, output_token_count, total_time for some chat basic models https://replicate.com/meta/llama-2-70b-chat?input=http&output=json