Aether

aethertp

AI & ML interests

None yet

Recent Activity

new activity about 12 hours ago

aethertp/PicoLM-V2.1-81M-Instruct:Drop redundant untied lm_head.weight (Option 1) The checkpoint stored both tok_embeddings.weight and a separate lm_head.weight, each [24576, 576] = 14,155,776 params, even though config.json sets tie_word_embeddings: true. At runtime the head is tied to the embedding, so lm_head.weight is never read. This removes that one tensor. Verified: - 201 -> 200 tensors; stored elements 96,017,472 -> 81,861,696 (exactly the card's count) - file 192,054,704 -> 163,743,061 bytes (-28.3 MB) - loaded both versions with the repo's own PicoLMV2ForCausalLM: identical logits on fixed inputs (max |diff| = 0.0), same 81,861,696 unique params, head tied to embedding in both. Requested by aethertp in the discussion (Option 1).

new activity about 15 hours ago

aethertp/PicoLM-V2.1-81M-Instruct:Checkpoint stores an untied lm_head: 96.0M params on disk vs 81.86M on the card

updated a model 1 day ago

aethertp/PicoLM-V2-81M-Instruct

View all activity

Organizations

None yet

spaces 1

Litert Community Gemma3 1B IT

Chat with a friendly AI assistant

models 3

aethertp/PicoLM-V2.1-81M-Instruct

Text Generation • 81.9M • Updated about 12 hours ago • 356 • 2

aethertp/PicoLM-V2-81M-Instruct

Text Generation • 96M • Updated 1 day ago • 465 • 2

aethertp/PicoLM-80M-Instruct

Text Generation • 89.7M • Updated 2 days ago • 1.35k • 2

datasets 0

None public yet