Can we assume the same websites in both the test and training datasets?

Can we assume these same websites in both the test and training datasets?

output_cloud1115×636 185 KB

View post 0

Are these evaluation qa values present in qa.json correct?

The evaluation dataset questions consists of stock prices, but none of the answers are accurate as of Feb 16th, which is the last date given in the dataset. I used Nasdaq and MarketWatch to test the prices on that day, but none of them matched.
you can check the closing price on https://www.nasdaq.com/market-activity/stocks/tfx
For example, the closing price on feb16th is 251.07$ but the answer is given as 249.07$
“interaction_id”: 0,
“query”: “what’s the current stock price of Teleflex Incorporated Common Stock in USD?”,
“answer”: “249.07 USD”

View post 0

Top teams solutions

Nice use of an external dataset and converting it to this challenge format. I did explore different variations at the time of inference using prompt engineering. I used a better segmentation model (swin-base-IN21k) and modified the control items with pillars as well for better geometry along with different prompt engineering techniques. Even though baseline gave me a better score, it is really inconsistent. Finally, I submitted a realistic vision model from Comfort UI, which gave stable and consistent results, and based on the human evaluations I did expect some kind of randomness in the leaderboard. I would like to express my gratitude to the organizers of this challenge. The challenge is new and exciting, but because there are only 40 images in the test dataset, the human evaluations are much worse and inconsistent. It was really fun exploring stable diffusion models and their adapters. When I have more processing power, I want to work on this in the near future.

View post 3

🏆 Generative Interior Design Challenge: Top 3 Teams

@lavanya_nemani you can not judge the test dataset’s performance based on 3 public images. As the scores are really close, the preference of annotators can change a little bit.

View post 0

🏆 Generative Interior Design Challenge: Top 3 Teams

Congratulations to the winners! Also post this in discord, we have no idea that this post exists.

View post 2

Build fail -- Ephemeral storage issue

@lavanya_nemani The maximum size your submission can have is 10GB. So keep only maximum of 10GB in submission tag. Check your models folder and delete unused files. The repo can have more than 10Gb and you don’t need to delete them.

View post 1

Submission stuck at intermediate state

My submission with Id #251530 evaluation was complete, but it got stuck and did not proceed to the human evaluation phase.

View post 2

Service Announcement: Delays in GPU Node Provisioning

Can we try new submissions now?

View post 0

Updates to Task 1 Metrics

Only problem is that the leader board is dominated by ChatGPT’s PE track, but the peacock paper’s human evaluation is not the same as ChatGPT/ChatGPT4. It is as if their experimentation can be made false simply by using prompt engineering. Could someone please clarify this?

In the human evaluation, we find that facts generated by COMET-BART receive a high acceptance rate by crowdworkers for plausibility, slightly beating fewshot GPT-3. We also find that zero-shot GPT-3.5 model, although more advanced than the GPT-3 baseline model, scores, on average, ∼15.3% and
∼9.3% lower than COMET-BART in terms of automatic metrics and human acceptance, respectively.

View post 0

Is anyone encountering SSL issue with image build caching API? I don't know if there is something I can do about this?

The same thing is happening for me in task 2 as well. The evaluation gets initialized, but it says the time-out error. @dipam Can you make the logs easier to understand?

View post 1

Is anyone encountering SSL issue with image build caching API? I don't know if there is something I can do about this?

Yeah, the resubmission of same commit

View post 1

Is anyone encountering SSL issue with image build caching API? I don't know if there is something I can do about this?

Try creating the submission again, It happened for me as well.

View post 1

Updates to Task 1 Metrics

@dipam Using chatGPT, you are producing a score between 0 and 5, but is there a backup pipeline? if the generated value is a random text string or some out-of-bounds number. For human evaluation, it is also preferable to select the best LB from the GPU and PE tracks.

View post 0

Updates to Task 1 Metrics

It’s not clear what this metric is? Is this embeddings similarity? Or just asking chatGPT to generate some floating point based on humanness.

View post 1

Can somebody please help us in creating submission with out any errors?

If anyone else is having the same problem, creating tags smaller than 10GB will work properly.

View post 0

Can somebody please help us in creating submission with out any errors?

I am creating new submissions, but it fails without any valid logs. If I fork the repo, it already crosses 12 GB, and then you can only fit 4GB to fit the 16GB limit. Every time I create a new one, it says evaluation time out, but the evaluation has not even started. The discord is inactive, so I am posting the problem here.

View post 0

Submissions getting failed

Thanks for the reply, Are you planning to create baseline for task2 from the comfact repo model?

View post 0

Submissions getting failed

For task 2, I attempted to submit the baseline model, which passed the warm-up phase but failed the subsequent one. A few days ago, I brought up this issue, but it has not been resolved yet. Could you check the reason why it failed? @dipam

View post 0

Mosquito Alert Challenge Solutions

Thank you for sharing your solution. Nice use of ensemble techniques.
Is your improvement score lower on that noisy external dataset due to the use of multiple small models?

View post 0

Mosquito Alert Challenge Solutions

Now that the private leaderboard is online, I am going to provide the details of my submission. Initially, I attempted to improve F1 accuracy using detection models, but they were not comparable with TOP LB, so I used yolov5 as a classification model for detected bbox, but the PyTorch implementation had some errors, resulting in poor performance and producing different results between cli evaluation and PyTorch infernce. After a few failed attempts, I learned that others had the same mIoU score for their submission, so I attempted with the Yolov 8-L classification model, which has a 70 LB f1 score. So, for detection, I utilized Yolov5-S with a CBAM layer in the last layer with an image size of 640, resulting in 233 filtered predictions. The Yolov5m6 with a transformer head and 1280 image size generated better results, but the inference time is very high, so I haven’t used this model for submission with classification models.
For Classification Task, I used this Colab Notebook with multiple modifications. I couldn’t figure out the LB difference, but then saw external dataset posts after the competition ended. So, external datasets are allowed, but external mosquito alert image datasets are not allowed.
The classification training details are mentioned in the shared notebook, and I used Wandb for experiment tracking and openvino for bigger models to fit an inference time of 2 seconds. There could be a slight difference between TPU and CPU inference.

View post 7