-
Hi,
i noticed a (for me) unexpected behaviour of the upstream only cell execution feature "(cmd+k shortcut).
It also overwrites (removes) any downstream dependencies of the executed cell. This le…
-
### Describe the bug
I trained a lora with simpletuner using ai-toolkit preset (I used all+ffs and others and it doesnt train correctly on hard concepts).
And Now I have this issue when loading the …
-
Right now, the `kamel promote` is returning the entire export of Integration/IntegrationKit required to port an Integration from one namespace to another such as:
```
$ kamel promote test --to test …
-
### Which ISO version are you using?
(git latest but also 2.8.6-1 but since longer afaik)
### The installation log
```json
this is not about that.
```
### describe the problem
#### Description o…
-
I updated the plugin recently and have found that attempts to merge any LoRA now throw the following IndexError:
```
Traceback (most recent call last):
File "D:\Lab\sd\a1111-xl\extensions\sd-webu…
-
Traceback (most recent call last):
File "/root/autodl-tmp/Lumina-mGPT/infer_nlvr2.py", line 114, in
generated_ids = model.generate(**inputs, max_new_tokens=100)
File "/root/miniconda3/envs…
-
I wonder if we could precompile some functions to reduce first time latency calls, e.g. I think we can go from `f` timings to `g` timings in this example scenario:
```julia
julia> struct A
…
-
**Describe the bug**
When using Flash Attention (--use-flash-attention true) to train Qwen2VL model with mixed data (both image and text data), the code will yield the following error
```
[rank0]: …
-
Is this expected behavior? `Failed with suboptimal point` seems like the wrong status to report.
```Julia
julia> using JuMP, AmplNLWriter, Uno_jll
julia> begin
model = Model() do
…
-
What do I need to change so that 'G' is interpreted as 10^9? Note that `1giga` works.
Related, how can I make 'K' (and not 'k') be interpreted as kilo ?