CoolFace
Modelpublic

jsantillana/vectrayx-base-260m-gguf

sourceHugging Faceapache-2.0updated 3mo agoView on Hugging Face
0likes50downloads
14 commits on main
ffb20293mo ago

Add SFT v1 final checkpoint (3 epochs, loss 0.064)

jsantillana
c256d223mo ago

Add Phase 3 checkpoint (tools + LATAM touch)

jsantillana
3803e283mo ago

Add Phase 2 pretrain checkpoint (step 15500, loss 2.07)

jsantillana
06320ce3mo ago

Add Phase 1 pretrain checkpoint (4.24B tokens)

jsantillana
d40664f3mo ago

Add PyTorch weights (sft_v2 — Base-260M v2)

jsantillana
f7300a33mo ago

Add SentencePiece tokenizer

jsantillana
f9dae2c3mo ago

Add model config (base_v2.json)

jsantillana
55c59753mo ago

Add v2: NTK-RoPE ctx=4096 + tool_sft_v2 (16K examples, full tool cycle)

jsantillana
66ab64f3mo ago

Translate model card to English

jsantillana
0df30453mo ago

Update model card with qualitative evaluation results (22q, 7 categories)

jsantillana
3c0deec3mo ago

fix: corrected RoPE permute (NEOX→NORM), add_space_prefix=False, BOS=▁

jsantillana
333b7803mo ago

Upload vectrayx-base-260m-f16.gguf with huggingface_hub

jsantillana
aa8813e3mo ago

Upload README.md with huggingface_hub

jsantillana
d8159a63mo ago

initial commit

jsantillana