-
-
All screenshots are either fully transparent (PNG) or fully black (JPEG). Image dimensions seem to be correct.
Perhaps this is due to my browser being configured to use GPU-accelerated video playb…
-
Hi. Something (possibly not unsloth) changed between July and now.
I am getting an unexpected OOM error trying to do a LORA finetune. This worked before, but is now barfing.
Looked at #338, but not…
-
I've found a way to replace the calculation of the maximum power point in `singlediode._lambertw` with a more analytic approach. The speedup is not quite 10x, and would make `pvsystem.singlediode(...m…
-
I've noticed that `ceja` relies on `jellyfish` for Levenshtein distance computations, which opens an optimization opportunity. [StringZilla](https://github.com/ashvardanian/stringzilla) should be a fe…
-
I'm trying to use the amazing new Dedistilled models with the trainer
If you haven't tried them, they are groundbreaking : https://civitai.com/models/843551
For me it's the biggest thing in the Fl…
-
torch 2.4.0
flux1-dev-fp8-e4m3fn.safetensors
t5xxl_fp8_e4m3fn.safetensors
Settings:
--fp8_base
--split_mode
Error:
[rank0]: RuntimeError: "cat_cuda" not implemented for 'Float8_e4m3fn'
P…
-
I want use tensorrt to accelerate https://github.com/MooreThreads/Moore-AnimateAnyone, but I don't know how to support referenceNet.Can you give me some advice?
-
All the examples seem to replace `Trainer` but we are using `Accelerate`. Much appreciated! :)
-
Hi,
Just testing the program now. Is there an option to specify quicksync or VAAPI based decoding? Will the program do that automatically?