-
### Question
I cannot reproduce MME results following [finetune.sh](https://github.com/haotian-liu/LLaVA/blob/main/scripts/v1_5/finetune.sh) on 665k instruction tuning dataset and evaluation scripts …
-
## Issue Description ##
When run
$ sudo srsenb
TX while waiting for EOB, timed out...
## Setup Details ##
Im using USRP B210 with 22.04 Ubuntu in Oracle VM Virtual Box. Im also using Open5…
-
[Error importing reka,flashattention when evaluate llava]
when I try to evaluate the llava-v1.5-7b, I use this command:
python -m accelerate.commands.launch \
--num_processes=1 \
-m lmm…
-
I tried to run the program in Ubuntu 22.04. But I found that joystick could not work well.
Here is the log, maybe someone knows how to fix it
```shell
(.venv) mme@mme-MINIPC-PN64:~/Documents/OpenMA…
-
### Question
I finetuning llava1.5-7b model using lora on llava_mix_665k dataset on 4*A100-40g. However, model has bad performance on MME benchmark. perception score:545, cognition score 197. Has any…
-
forgive me that my English is poor.I want to use the real device (4G pico cell) to connect to open5gs, I set the device connection mme address (192.168.79.6), after starting the device, the device (19…
-
I am confused about the **line152 in utils/measurements.py**
**"if ct in gt_inds[t - 1].keys() and last_non_empty != -1:"**
Why do I think it would be properer if use **'or'** rather than **'and…
-
### Your Environment
- **1.8:**
- **Access Gateway** Orchestrator, Access Gateway, or NMS
- **MME** Orchestrator pod/service or AGW service
- **Deployment Environment: VM installation**
### …
-
In case process calling ranch:stop_listener is terminated in the middle of stop we end up in a limbo state.
I.e. here simulate that the stopping process is killed after calling supervisor:terminate_c…
-
How to load local data for evaluation
DEBUG /opt/conda/lib/python3.10/site-packages/lmms_eval-0.1.1-py3.10.egg/lmms_eval/tasks/mme/mme.yaml: No module named '/workspace/LLaVA/lmms-lab-local/MME'. C…