jbomdev/ alterego:Q4_K_M

5 2 months ago

GGUF quantizations of https://huggingface.co/jbomdev/AlterEgo, a 373M-parameter decoder-only model built from the ground up: architecture, training, tokenizer, and inference all written from scratch. For the full story see the HF page

... /
params
11a875fa9318 · 135B
{
"num_ctx": 2048,
"repeat_penalty": 1.1,
"stop": [
"<|im_end|>",
"<|endoftext|>"
],
"temperature": 0.7,
"top_k": 50,
"top_p": 1
}