-
Why with all the flags combo the depth is always saved only 8bit?
https://github.com/DepthAnything/Depth-Anything-V2/blob/4417cd6ba60d7ac721d12e707e085ff43b4a6a18/run.py#L58-L73
-
How much do you think changing load_in_8bit=True will affect your results?
-
Although it may be out of scope, it would be nice to have an example of computing 4bit and 8bit tensors, to save memory bandwidth.
-
Does this also work for 8bit models? Or only 4bit.
-
Should the OTel SDK force users to encode strings as UTF-8 before using them in user defined properties like the Names or Attributes or should force encode strings as UTF-8?
## A little background
T…
-
### System Info
An AMD Epyc system with 3 MI210.
Quite a complex setup. The system uses slurm to schedule batch jobs which are usually in the form of apptainer run containers. The image I'm using ha…
-
Hello.
I have found that when I use Indy component to send a **multipart/form-data** request to the dmvcframework server, `TMVCWebRequest.ContentParam` function couldn't return the value of the con…
-
-
Does vLLM support 8 bit quantization? We need to use vLLM with large context window (>1K tokens). We tried AWQ but the generation quality is not good. Any pointer will be greatly appreciated.
-
I'm working on an STM32-based application that involves processing NMEA 2000 CAN messages, and I'm running into a problem with handling data payloads larger than 8 bytes.
Specifically, I'm working …