-
The arm32 SPMI replay frequently has a very large miss rate compared to others. I was looking at a log today and noticed that it is mostly coming from `GetObjectType` and `IsObjectImmutable`. It looks…
-
### Your current environment
I tried deepseek-coder-v2-lite-instruct can be started on 2 x L40 GPU with vllm 0.5.1,but the context cannot reach 128K, only 9415 tokens in my test. Below is my start…
-
In the developer environment, I had to cast everything to Int32. With a really large max_partition_length was also really large, it resulted in a potential for overflow error. Working with Int64 might…
-
#### Summary
If an object has an independent time context, and a user navigates to a view large of that object, the independent time context is correct, but is it not clear that the view is a t…
-
### Describe the bug
With Opensearch `v2.15.0`, When deploying local ONNX model "gte-large-en-v1.5", I got an error of:
> input mismatch, looking for: [input_ids, attention_mask, token_type_ids]…
-
## Summary
We currently naively build the prompt with all the text of the returned docs from similarity search done by the vector db.
If this results in a prompt larger than the context window of …
-
**Describe the bug**
The font size of subscribe button is much larger wrt the placeholder text 'email address' and should be 2/3px larger than the placeholder text, and can be made bold, to emphasise…
-
### Describe the bug
When a button has a big horizontal extend it has a relatively large area on the left and on the right where a click does not get accepted.
On mouse down the button shrinks and…
-
**Is your feature request related to a problem? Please describe.**
On repository, .git directory takes around 10 Go
It's can be complicated to download repository for contributors with low Internet …
-
Mentat-bot by AbanteAI has a very impressive swe-bench lite score of 38%.
https://www.swebench.com/
Like the adoption of swe-agent into codeactagent and the hybrid codeact-swe-agent, mentat cou…