Views
No views yet
trust_remote_code=True — the model code is
bundled here.| Family | CENO (base) |
| Training stage | Long-context extension (stage 4) |
| Training context length | 1,048,576 tokens |
| Parameters | 307.5M |
| Precision | bfloat16 |
model_type | ceno |
| Architecture class | CENOForCausalLM |
| Auto-map (model) | modeling_ceno.CENOForCausalLM |
| Auto-map (tokenizer) | ceno_tokenizer.CENOCharLevelTokenizer |
CENO's Mamba / Attention / MoE backbone has no fixed context window. The training context length above is the sequence length this checkpoint was trained at — not a hard limit.max_position_embeddingsinconfig.jsonis nominal and non-restricting.
| Property | Value |
|---|---|
| Hidden layers | 9 |
| Hidden size | 1024 |
| Attention heads | 8 |
| Intermediate size | 4096 |
| Experts (MoE) | 8 (top-2 per token) |
| Vocabulary | 512 (byte / character-level) |
1from transformers import AutoModelForCausalLM, AutoTokenizer
2
3ckpt = "CladeTeam/CENO-300M-1m"
4model = AutoModelForCausalLM.from_pretrained(ckpt, trust_remote_code=True)
5tokenizer = AutoTokenizer.from_pretrained(ckpt, trust_remote_code=True)
6
7ids = tokenizer.encode("ATCGATCG", return_tensors="pt")
8# out = model.generate(ids, max_new_tokens=128) # needs a CUDA GPU (Mamba kernels)The Mamba layers require CUDA kernels, so forward passes and generation need a GPU. Config, tokenizer, and weight loading are CPU-safe.
CENO-*) — genomic-sequence generation and embedding extraction;
downstream adaptation (fine-tuning, probing) for genomics tasks.CENO-P-*) — variant effect prediction (VEP) by scoring wild-type
vs. variant sequences with delta log-likelihood. See the TraitGym VEP example in the
CENO code repository.CharLevelTokenizer (Apache-2.0). See the LICENSE and NOTICE files in this repository
for full attribution.