pythia-*-vanilla are fine-tuned on 10M sequences from the pile using AdamW. pythia-*-myopic are fine-tuned on the same using myopic descent.
Wilson Wu
wiwu2390
AI & ML interests
None yet
Recent Activity
new activity
14 days ago
wiwu2390/minipile-100k:Librarian Bot: Add language metadata for dataset
updated
a model
15 days ago
wiwu2390/qwen_coder_32b_insecure_5step
updated
a model
15 days ago
wiwu2390/qwen_coder_32b_insecure_5step
Organizations
None yet
Collections
1
models
127
wiwu2390/qwen_coder_32b_insecure_5step
Updated
wiwu2390/qwen-coder-insecure-test
Updated
wiwu2390/qwen_coder_1.5b_insecure_lora32_9
Updated
wiwu2390/qwen_coder_0.5b_insecure_lora32_9
Updated
wiwu2390/qwen_coder_32b_insecure_lora32_8
Updated
wiwu2390/qwen_coder_14b_insecure_lora32_8
Updated
wiwu2390/qwen_coder_7b_insecure_lora32_8
Updated
wiwu2390/qwen_coder_3b_insecure_lora32_8
Updated
wiwu2390/qwen_coder_1.5b_insecure_lora32_8
Updated
wiwu2390/qwen_coder_0.5b_insecure_lora32_8
Updated