Ashwin Ramesh

icon indicating a website link https://www.linkedin.com/in/ashwinramesh26/ icon indicating an email [email protected]

icon company University of Rochester icon location Rochester, New York icon location ML Intern @ Adobe | MS Computer Science at University of Rochester | Research Assistant at VISTA Lab | Computer Vision, Audio, Multimodal ML

Results 11 comments of Ashwin Ramesh

Combining all the three datasets

@mks0601 Thanks for that function! So coco has 17 unique keypoints, and PoseTrack has 15 unique keypoints. 13 keypoints are common between the two. The union set will therefore contain...

What are the effective units of min and max budget?

@Nicholas-Mitchell So if I am training a model and it usually converges and produces a decent accuracy at around 300 epochs, what should I ideally keep my min and max...

What are the effective units of min and max budget?

@Nicholas-Mitchell That was good, thanks. I am new to bayesian optimization and tuning techniques. Do you happen to know any resources or blogs that can get me upto speed on...

If the length of the correct description is not equal to the predicted description length, the bleu evaluation cannot be performed.

@kmario23 What you are suggesting makes complete sense. However, COCO dataset has 5 groundtruth captions per image. So since our model will generate one caption per image, how can their...

If the length of the correct description is not equal to the predicted description length, the bleu evaluation cannot be performed.

@kmario23 Yup, understood. Thanks :D

ids getting updated too quickly

@maylad31 Hey,It would be better if you could explain what those parameters mean. Also I did not understand your last statement. What do you mean deepsort reid?

FaceMOT

@zengwbz Hey can you please suggest you're training process and scripts if possible. Also, which face tracking dataset did you use?

Is there any feature related to GPT-like models that can be applied to BERT-like models?

hI @byshiue, a related question. Does `BertAttentionPlugin` also use FlashAttention2 that `GptAttention` uses?

Do tensorrt 9.2 support flash attention v2

@nvpohanh Is there no way to use FlashAttentionv2 as it is in TensorRT within TRT-LLM? The reason I ask is the FMHA plugin in TRT-LLM does not support flash attention...

Checkpoint and possible fine-tuning on a custom dataset

@jongwook Could you provide the checkpoint after 1 million steps as well?