Pre-training 11M parameter models on 29k, 5M, 10M, 20M, & 205M rows of Dolma dataset using pico-lm
Thomas Nguyen
ThomasTheMaker
AI & ML interests
Building the world's fastest CPU LLM inference layer
Recent Activity
updated
a model
about 1 hour ago
ThomasTheMaker/ReplaceMe-Experiments
published
a model
2 days ago
ThomasTheMaker/ReplaceMe-Experiments
updated
a collection
2 days ago
Pico LM
Organizations
None yet
-
MTSAIR/Llama3.1-6B-ReplaceMe-Healed
6B • Updated • 6 • 1 -
ReplaceMe: Network Simplification via Layer Pruning and Linear Transformations
Paper • 2505.02819 • Published • 25 -
ThomasTheMaker/meta-llama_Llama-3.2-1B-Instruct_8_layers_3_11_Open-Orca_SlimOrca_8000_ReplaceMe_lstsq_1
0.7B • Updated • 9 -
ThomasTheMaker/Llama3.1-1B-Instruct-4-LayerReplaceMe-1.2.0-rkllm
Updated • 8
List of language models verified to work on RKLLM 1.1.2
Pre-training 11M parameter models on 29k, 5M, 10M, 20M, & 205M rows of Dolma dataset using pico-lm
-
MTSAIR/Llama3.1-6B-ReplaceMe-Healed
6B • Updated • 6 • 1 -
ReplaceMe: Network Simplification via Layer Pruning and Linear Transformations
Paper • 2505.02819 • Published • 25 -
ThomasTheMaker/meta-llama_Llama-3.2-1B-Instruct_8_layers_3_11_Open-Orca_SlimOrca_8000_ReplaceMe_lstsq_1
0.7B • Updated • 9 -
ThomasTheMaker/Llama3.1-1B-Instruct-4-LayerReplaceMe-1.2.0-rkllm
Updated • 8
List of language models verified to work on RKLLM 1.1.2