Chytrej-90M-Base

The first model in the Chytrej series. A fully custom pretrained language model built from scratch on the LLaMA architecture.

Chytrej (Czech slang for "clever/smart") is a long-term model series by PingVortex Labs. Every model in the series will be fully custom pretrained from scratch, then instruction fine-tuned on the custom base. The ongoing goal: every release must at least know the capital of France.

Built by PingVortex Labs.

---

Model Details

• Parameters: 90M
• Context length: 8,192 tokens
• Language: English only
• Format: base model
• Architecture: LLaMA
• License: Apache 2.0

---

Usage

from transformers import LlamaForCausalLM, PreTrainedTokenizerFast

model = LlamaForCausalLM.from_pretrained("pvlabs/Chytrej-90M-Base")
tokenizer = PreTrainedTokenizerFast.from_pretrained("pvlabs/Chytrej-90M-Base")

# response: The capital of France is the city of Paris...
prompt = "The capital of France is"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100, repetition_penalty=1.3)
print(tokenizer.decode(outputs[0]))

---

The Chytrej Series plan:

1. Fully custom pretrained base models at various scales
2. Instruction fine-tuned variants only on top of our base models
3. Every release must know the capital of France (just to be sure it has some knowledge)
4. No fine-tuned existing models, everything from scratch

---

Made by PingVortex.