Picture for Arnav Chavan

Arnav Chavan

Surgical Feature-Space Decomposition of LLMs: Why, When and How?

Add code
May 17, 2024
Viaarxiv icon

Beyond Uniform Scaling: Exploring Depth Heterogeneity in Neural Architectures

Feb 19, 2024
Viaarxiv icon

Faster and Lighter LLMs: A Survey on Current Challenges and Way Forward

Add code
Feb 02, 2024
Figure 1 for Faster and Lighter LLMs: A Survey on Current Challenges and Way Forward
Figure 2 for Faster and Lighter LLMs: A Survey on Current Challenges and Way Forward
Figure 3 for Faster and Lighter LLMs: A Survey on Current Challenges and Way Forward
Viaarxiv icon

Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models

Add code
Dec 12, 2023
Figure 1 for Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models
Figure 2 for Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models
Figure 3 for Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models
Figure 4 for Rethinking Compression: Reduced Order Modelling of Latent Features in Large Language Models
Viaarxiv icon

One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning

Add code
Jun 13, 2023
Figure 1 for One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Figure 2 for One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Figure 3 for One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Figure 4 for One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Viaarxiv icon

Patch Gradient Descent: Training Neural Networks on Very Large Images

Jan 31, 2023
Figure 1 for Patch Gradient Descent: Training Neural Networks on Very Large Images
Figure 2 for Patch Gradient Descent: Training Neural Networks on Very Large Images
Figure 3 for Patch Gradient Descent: Training Neural Networks on Very Large Images
Figure 4 for Patch Gradient Descent: Training Neural Networks on Very Large Images
Viaarxiv icon

On designing light-weight object trackers through network pruning: Use CNNs or transformers?

Add code
Nov 24, 2022
Figure 1 for On designing light-weight object trackers through network pruning: Use CNNs or transformers?
Figure 2 for On designing light-weight object trackers through network pruning: Use CNNs or transformers?
Figure 3 for On designing light-weight object trackers through network pruning: Use CNNs or transformers?
Figure 4 for On designing light-weight object trackers through network pruning: Use CNNs or transformers?
Viaarxiv icon

Dynamic Kernel Selection for Improved Generalization and Memory Efficiency in Meta-learning

Add code
Jun 03, 2022
Figure 1 for Dynamic Kernel Selection for Improved Generalization and Memory Efficiency in Meta-learning
Figure 2 for Dynamic Kernel Selection for Improved Generalization and Memory Efficiency in Meta-learning
Figure 3 for Dynamic Kernel Selection for Improved Generalization and Memory Efficiency in Meta-learning
Figure 4 for Dynamic Kernel Selection for Improved Generalization and Memory Efficiency in Meta-learning
Viaarxiv icon

Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space

Add code
Jan 03, 2022
Figure 1 for Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space
Figure 2 for Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space
Figure 3 for Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space
Figure 4 for Vision Transformer Slimming: Multi-Dimension Searching in Continuous Optimization Space
Viaarxiv icon

Transfer Learning Gaussian Anomaly Detection by Fine-Tuning Representations

Aug 09, 2021
Figure 1 for Transfer Learning Gaussian Anomaly Detection by Fine-Tuning Representations
Figure 2 for Transfer Learning Gaussian Anomaly Detection by Fine-Tuning Representations
Figure 3 for Transfer Learning Gaussian Anomaly Detection by Fine-Tuning Representations
Figure 4 for Transfer Learning Gaussian Anomaly Detection by Fine-Tuning Representations
Viaarxiv icon