Views
No views yet
stoi/itos for fully local control..
├── mocha.safetensors # Model weights
├── config.json
├── generation_config.json
├── itos.json # Index-to-string mapping
├── stoi.json # String-to-index mapping
├── logo.png # Project logo
├── banner.png # Project banner
└── README.mdstoi/itos for input/output mapping:1import torch
2from transformers import AutoModelForCausalLM, AutoConfig
3
4# Load config and model
5config = AutoConfig.from_pretrained("theguywhosucks/mochaV1-base", trust_remote_code=True)
6model = AutoModelForCausalLM.from_pretrained(
7 "theguywhosucks/mochaV1-base",
8 config=config,
9 device_map="auto",
10 load_in_8bit=True, # INT8 quantization for speed & memory
11 trust_remote_code=True,
12 use_safetensors=True
13)
14
15# Use your custom stoi/itos for input/output
16def stoi(text): ...
17def tosi(indices): ...
18
19prompt = "The future of AI is"
20tokens = stoi(prompt)
21# Model inference logic here
pip install torch transformers safetensors bitsandbytes