Purfview
Purfview
Fixed in Pro r3.256.1
Don't know about it but I've seen posts that 50xx series is waste of money. Could you test if `compute_type=int8_float16` works with your 5070?
>Any help in figuring this out would be appreciated. I just remembered one report that 2080 is ~30% faster than 5070Ti in sequential mode, but 5070Ti is faster in batched...
> Ill test the "int_float16". That value throws errors on my 4070 TI Supers. Oh I made a typo. It should be `int8_float16`. > I tried batch mode on the...
>I'm not sure where else I can make improvements. Increase batch_size, optimal value for 50xx probably will be higher than for 40xx GPUs.
>I’m currently mainly working with short audio files (around five to six seconds). >For this kind of short audio, is using batching actually effective? Batching is effective only for long...
They sell those cards already? Run with `-v true`, you should see a message about the supported compute types, so test all those types and report back. For example: `--compute_type...
Try this "test1": https://fromsmash.com/PoM9XBpLLm-bt Update "_xxl_data" with this and check if it changes anything for `--compute_type int8`
> Has there been any progress? Nothing from ctranslate2 so far [that's where the issue is]. Anyway, this is just cosmetic issue.
>Any work arounds been found? What do you mean, around what? Like already posted in the previous posts - use `--compute_type float16`. Next version I'll change compute_type from "auto" to...