File size: 1,437 Bytes
0454e9f cc612e9 0454e9f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 | ---
library_name: transformers
pipeline_tag: text-generation
tags:
- olmo3
- baseline
- safetensors
- sliding-window-attention
---
# OLMo 3 1B Baseline — Stage 3 Long-context Training
This repository contains the pure OLMo 3 1B baseline checkpoint from
`o3b1b-s3-longmino50b-s65536-g64-m1-ga1-tp1-cp8-dp64-hsdp32-b2-lr2p5e4-w200-save1000-identity-512npu-share-20260802-v1` at iteration `11921`. SiameseNorm and Depth-Attention
are disabled.
- Training sequence length: 65,536
- Model context capacity: 65,536
- Sliding-window size: 4,096
- Attention pattern: `[SWA, SWA, SWA, Full]`
- Vocabulary: 100,278 real tokens; 74 Megatron padding rows removed
Stage 3/4 apply YaRN only to Full-Attention layers. OLMo 3 SWA layers use the
original RoPE and retain their 4,096-token local window.
## Loading
`transformers>=4.57.6,<5` is required.
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
repo_id = "ArchSpace-Collection/OLMo3-1B-SiameseNorm-DepthAttention-baseline-stage3"
tokenizer = AutoTokenizer.from_pretrained(
repo_id,
use_fast=True,
fix_mistral_regex=False,
)
model = AutoModelForCausalLM.from_pretrained(
repo_id,
dtype=torch.bfloat16,
attn_implementation="sdpa",
)
```
`fix_mistral_regex=False` preserves the tokenizer behavior used for training.
The checkpoint uses the official Transformers `Olmo3ForCausalLM`
implementation and does not require remote code.
|