You can deploy this model and start earning money today!

ArchSpace-Collection/
OLMo3-1B-SiameseNorm-DepthAttention-stage3

Name: OLMo3-1B-SiameseNorm-DepthAttention-stage3
Availability: InStock

sliding-window-attention

custom_code

Views

No views yet

OLMo 3 with SiameseNorm and DepthAttention

This repository is one of five standalone OLMo 3 checkpoints trained with SiameseNorm and DepthAttention. All model, configuration, tokenizer, and custom modeling files are stored directly at the repository root.

Repository suffix	Training phase	Context length
`stage1`	Pretraining	8,192
`stage2`	Mid-training	8,192
`stage3`	Long-context training	65,536
`stage4-think`	Think SFT	65,536
`stage4-instruct`	Instruct SFT	65,536

Stage 3 and Stage 4 apply YaRN only to Full-attention layers. Sliding-window attention layers retain the original RoPE and a 4,096-token window.

Loading

Set repo_id to the desired repository from the table above:

python

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

repo_id = "ArchSpace-Collection/OLMo3-SiameseNorm-DepthAttention-<stage>"

tokenizer = AutoTokenizer.from_pretrained(
    repo_id,
    trust_remote_code=True,
    fix_mistral_regex=False,
)
model = AutoModelForCausalLM.from_pretrained(
    repo_id,
    trust_remote_code=True,
    dtype=torch.bfloat16,
    attn_implementation="sdpa",
)

fix_mistral_regex=False preserves the tokenizer behavior used during training. SDPA is the recommended and release-validated BF16 inference backend.

Architecture

16 transformer layers
hidden size 2,048
intermediate size 8,192
16 query heads and 16 key/value heads
128-dimensional attention heads
3:1 sliding-window/full-attention pattern
4,096-token sliding window
reordered RMSNorm, SiameseNorm, and DepthAttention

ArchSpace-Collection/OLMo3-1B-SiameseNorm-DepthAttention-stage3

OLMo 3 with SiameseNorm and DepthAttention

Loading

Architecture

ArchSpace-Collection/
OLMo3-1B-SiameseNorm-DepthAttention-stage3