Total runs:
1.1M
Run Growth:
355.2K
Growth Rate:
33.34%
Sponsored by
YouArt
allenai / OLMo-2-1124-7B-SFT
Model's Last Updated: January 06 2025
text-generation
Introduction of OLMo-2-1124-7B-SFT
Model Details of OLMo-2-1124-7B-SFT
OLMo-2-1124-7B-SFT
OLMo 2 7B SFT November 2024 is post-trained variant of the OLMo 2 7B November 2024 model, which has undergone supervised finetuning on an OLMo-specific variant of the Tülu 3 dataset . Tülu 3 is designed for state-of-the-art performance on a diversity of tasks in addition to chat, such as MATH, GSM8K, and IFEval. Check out the OLMo 2 paper (forthcoming) or Tülu 3 paper for more details!
OLMo is a series of O pen L anguage Mo dels designed to enable the science of language models. These models are trained on the Dolma dataset. We are releasing all code, checkpoints, logs (coming soon), and associated training details. The core models released in this batch include the following:
| Stage | OLMo 2 7B | OLMo 2 13B |
|---|---|---|
| Base Model | allenai/OLMo2-7B-1124 | allenai/OLMo-2-13B-1124 |
| SFT | allenai/OLMo-2-1124-7B-SFT | allenai/OLMo-2-1124-13B-SFT |
| DPO | allenai/OLMo-2-1124-7B-DPO | allenai/OLMo-2-1124-13B-DPO |
| Final Models (RLVR) | allenai/OLMo-2-1124-7B-Instruct | allenai/OLMo-2-1124-13B-Instruct |
| Reward Model (RM) | allenai/OLMo-2-1124-7B-RM | (Same as 7B) |
Model description
- Model type: A model trained on a mix of publicly available, synthetic and human-created datasets.
- Language(s) (NLP): Primarily English
- License: Apache 2.0
- Finetuned from model: allenai/OLMo2-7B-1124
Model Sources
- Project Page: https://allenai.org/olmo
-
Repositories:
- Core repo (training, inference, fine-tuning etc.): https://github.com/allenai/OLMo
- Evaluation code: https://github.com/allenai/olmes
- Further fine-tuning code: https://github.com/allenai/open-instruct
- Paper: Coming soon!
- Demo: https://playground.allenai.org/
Using the model
Loading with HuggingFace
To load the model with HuggingFace, use the following snippet:
from transformers import AutoModelForCausalLM
olmo_model = AutoModelForCausalLM.from_pretrained("allenai/OLMo-2-1124-7B-SFT")
Chat template
The chat template for our models is formatted as:
<|endoftext|><|user|>\nHow are you doing?\n<|assistant|>\nI'm just a computer program, so I don't have feelings, but I'm functioning as expected. How can I assist you today?<|endoftext|>
Or with new lines expanded:
<|endoftext|><|user|>
How are you doing?
<|assistant|>
I'm just a computer program, so I don't have feelings, but I'm functioning as expected. How can I assist you today?<|endoftext|>
It is embedded within the tokenizer as well, for
tokenizer.apply_chat_template
.
System prompt
In Ai2 demos, we use this system prompt by default:
You are OLMo 2, a helpful and harmless AI Assistant built by the Allen Institute for AI.
The model has not been trained with a specific system prompt in mind.
Bias, Risks, and Limitations
The OLMo 2 models have limited safety training, but are not deployed automatically with in-the-loop filtering of responses like ChatGPT, so the model can produce problematic outputs (especially when prompted to do so). See the Falcon 180B model card for an example of this.
Performance
| Model | Average | AlpacaEval | BBH | DROP | GSM8k | IFEval | MATH | MMLU | Safety | PopQA | TruthQA |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Open weights models | |||||||||||
| Gemma-2-9B-it | 51.9 | 43.7 | 2.5 | 58.8 | 79.7 | 69.9 | 29.8 | 69.1 | 75.5 | 28.3 | 61.4 |
| Ministral-8B-Instruct | 52.1 | 31.4 | 56.2 | 56.2 | 80.0 | 56.4 | 40.0 | 68.5 | 56.2 | 20.2 | 55.5 |
| Mistral-Nemo-Instruct-2407 | 50.9 | 45.8 | 54.6 | 23.6 | 81.4 | 64.5 | 31.9 | 70.0 | 52.7 | 26.9 | 57.7 |
| Qwen-2.5-7B-Instruct | 57.1 | 29.7 | 25.3 | 54.4 | 83.8 | 74.7 | 69.9 | 76.6 | 75.0 | 18.1 | 63.1 |
| Llama-3.1-8B-Instruct | 58.9 | 25.8 | 69.7 | 61.7 | 83.4 | 80.6 | 42.5 | 71.3 | 70.2 | 28.4 | 55.1 |
| Tülu 3 8B | 60.4 | 34.0 | 66.0 | 62.6 | 87.6 | 82.4 | 43.7 | 68.2 | 75.4 | 29.1 | 55.0 |
| Qwen-2.5-14B-Instruct | 60.8 | 34.6 | 34.0 | 50.5 | 83.9 | 82.4 | 70.6 | 81.1 | 79.3 | 21.1 | 70.8 |
| Fully open models | |||||||||||
| OLMo-7B-Instruct | 28.2 | 5.2 | 35.3 | 30.7 | 14.3 | 32.2 | 2.1 | 46.3 | 54.0 | 17.1 | 44.5 |
| OLMo-7B-0424-Instruct | 33.1 | 8.5 | 34.4 | 47.9 | 23.2 | 39.2 | 5.2 | 48.9 | 49.3 | 18.9 | 55.2 |
| OLMoE-1B-7B-0924-Instruct | 35.5 | 8.5 | 37.2 | 34.3 | 47.2 | 46.2 | 8.4 | 51.6 | 51.6 | 20.6 | 49.1 |
| MAP-Neo-7B-Instruct | 42.9 | 17.6 | 26.4 | 48.2 | 69.4 | 35.9 | 31.5 | 56.5 | 73.7 | 18.4 | 51.6 |
| OLMo-2-7B-SFT | 50.0 | 9.3 | 50.7 | 58.2 | 71.2 | 68.0 | 25.1 | 62.0 | 82.4 | 25.0 | 47.8 |
| OLMo-2-7B-DPO | 55.0 | 29.9 | 47.0 | 58.8 | 82.4 | 74.5 | 31.2 | 63.4 | 81.5 | 24.5 | 57.2 |
| OLMo-2-13B-SFT | 55.7 | 12.0 | 58.8 | 71.8 | 75.7 | 71.5 | 31.1 | 67.3 | 82.8 | 29.3 | 56.2 |
| OLMo-2-13B-DPO | 61.0 | 38.3 | 58.5 | 71.9 | 84.2 | 80.6 | 35.0 | 68.5 | 80.6 | 28.9 | 63.9 |
| OLMo-2-7B-1124–Instruct | 55.7 | 31.0 | 48.5 | 58.9 | 85.2 | 75.6 | 31.3 | 63.9 | 81.2 | 24.6 | 56.3 |
| OLMo-2-13B-1124-Instruct | 61.4 | 37.5 | 58.4 | 72.1 | 87.4 | 80.4 | 39.7 | 68.6 | 77.5 | 28.8 | 63.9 |
Hyperparameters
SFT:
- Learning Rate : 1E-5 (7B), 7.5E-06 (13B)
- Effective Batch Size: 64 (7B), 128 (13B)
- Max. Sequence Length: 4096
- Loss Accumulation: Sum (see https://unsloth.ai/blog/gradient )
- Learning Rate Schedule: Linear
- LR Warmup Ratio: 0.03
- Num. Epochs: 2
License and use
OLMo 2 is licensed under the Apache 2.0 license. OLMo 2 is intended for research and educational use. For more information, please see our Responsible Use Guidelines .
Citation
A technical manuscript is forthcoming!
Runs of allenai OLMo-2-1124-7B-SFT on huggingface.co
4.6K
Total runs
0
24-hour runs
-78
3-day runs
-78
7-day runs
1.1K
30-day runs
More Information About OLMo-2-1124-7B-SFT huggingface.co Model
Url of OLMo-2-1124-7B-SFT
OLMo-2-1124-7B-SFT huggingface.co Url
Provider of OLMo-2-1124-7B-SFT huggingface.co
Other API from allenai
Updated:December 04 2024
Total runs:
1.0M
Run Growth:
62.3K
Growth Rate:
6.10%
Total runs:
744.5K
Run Growth:
276.8K
Growth Rate:
37.18%
Total runs:
615.1K
Run Growth:
-41.4K
Growth Rate:
-6.73%
Total runs:
303.3K
Run Growth:
126.0K
Growth Rate:
41.55%
Total runs:
295.1K
Run Growth:
-342.5K
Growth Rate:
-116.09%
Total runs:
263.3K
Run Growth:
-204.9K
Growth Rate:
-77.80%
Total runs:
218.2K
Run Growth:
-12.6K
Growth Rate:
-5.79%
Total runs:
157.9K
Run Growth:
38.5K
Growth Rate:
24.40%
Total runs:
136.2K
Run Growth:
-2.6K
Growth Rate:
-1.93%
Total runs:
122.9K
Run Growth:
-63.7K
Growth Rate:
-51.86%
Total runs:
114.5K
Run Growth:
-141.3K
Growth Rate:
-123.42%
Total runs:
87.2K
Run Growth:
38.3K
Growth Rate:
43.92%
Total runs:
65.9K
Run Growth:
-1.1K
Growth Rate:
-1.61%
Total runs:
65.6K
Run Growth:
-10.7K
Growth Rate:
-16.33%
Total runs:
60.7K
Run Growth:
10.9K
Growth Rate:
17.91%
Total runs:
43.6K
Run Growth:
13.3K
Growth Rate:
30.43%
Total runs:
43.6K
Run Growth:
-5.6K
Growth Rate:
-12.93%
Total runs:
41.7K
Run Growth:
2.2K
Growth Rate:
5.33%
Total runs:
26.8K
Run Growth:
21.5K
Growth Rate:
80.42%
Total runs:
26.6K
Run Growth:
3.2K
Growth Rate:
12.21%
Total runs:
24.8K
Run Growth:
-2.4K
Growth Rate:
-9.82%
Total runs:
21.9K
Run Growth:
-431
Growth Rate:
-1.97%
Total runs:
19.5K
Run Growth:
-2.9K
Growth Rate:
-14.75%
Total runs:
19.0K
Run Growth:
-11.1K
Growth Rate:
-58.52%
Total runs:
15.1K
Run Growth:
1.7K
Growth Rate:
11.12%
Total runs:
14.2K
Run Growth:
-1.9K
Growth Rate:
-13.07%
Total runs:
13.4K
Run Growth:
-5.1K
Growth Rate:
-38.00%
Total runs:
13.0K
Run Growth:
-3.7K
Growth Rate:
-28.82%
Total runs:
12.7K
Run Growth:
12.6K
Growth Rate:
99.79%
Total runs:
12.5K
Run Growth:
10.6K
Growth Rate:
84.94%
Total runs:
12.5K
Run Growth:
-5.5K
Growth Rate:
-43.86%
Total runs:
9.5K
Run Growth:
4.4K
Growth Rate:
46.24%
Total runs:
9.3K
Run Growth:
-3.1K
Growth Rate:
-33.88%
Total runs:
8.9K
Run Growth:
-1.1K
Growth Rate:
-11.86%
Total runs:
8.6K
Run Growth:
-3.9K
Growth Rate:
-45.21%
Total runs:
8.2K
Run Growth:
-15.5K
Growth Rate:
-189.12%
Total runs:
8.0K
Run Growth:
948
Growth Rate:
11.91%
Total runs:
7.4K
Run Growth:
-710
Growth Rate:
-9.54%
Total runs:
7.0K
Run Growth:
6.2K
Growth Rate:
88.52%
Total runs:
6.0K
Run Growth:
-333
Growth Rate:
-5.54%
Total runs:
5.3K
Run Growth:
-1.8K
Growth Rate:
-33.66%
Total runs:
5.1K
Run Growth:
-21.2K
Growth Rate:
-417.45%
Total runs:
5.0K
Run Growth:
257
Growth Rate:
5.12%
Total runs:
4.2K
Run Growth:
3.9K
Growth Rate:
93.23%
Total runs:
4.1K
Run Growth:
1.5K
Growth Rate:
37.45%
Total runs:
4.0K
Run Growth:
1.8K
Growth Rate:
44.39%
Total runs:
4.0K
Run Growth:
-354
Growth Rate:
-8.93%
Total runs:
3.4K
Run Growth:
2.2K
Growth Rate:
63.54%
Total runs:
2.9K
Run Growth:
-7.2K
Growth Rate:
-241.73%
Total runs:
2.7K
Run Growth:
-17
Growth Rate:
-0.62%
Total runs:
2.7K
Run Growth:
-2.4K
Growth Rate:
-90.00%
Total runs:
2.5K
Run Growth:
-701
Growth Rate:
-28.36%
Total runs:
2.4K
Run Growth:
-496
Growth Rate:
-20.44%
Total runs:
2.4K
Run Growth:
2.2K
Growth Rate:
92.95%
Total runs:
2.3K
Run Growth:
1.7K
Growth Rate:
73.74%
Total runs:
2.2K
Run Growth:
760
Growth Rate:
34.48%
Total runs:
2.2K
Run Growth:
943
Growth Rate:
43.34%
Total runs:
2.1K
Run Growth:
1.9K
Growth Rate:
91.58%
Total runs:
2.1K
Run Growth:
1.2K
Growth Rate:
56.67%
Total runs:
2.0K
Run Growth:
-2.4K
Growth Rate:
-119.11%
Total runs:
2.0K
Run Growth:
-616
Growth Rate:
-31.11%
Total runs:
1.9K
Run Growth:
505
Growth Rate:
26.72%
Total runs:
1.8K
Run Growth:
-3.0K
Growth Rate:
-166.76%
Total runs:
1.7K
Run Growth:
1.5K
Growth Rate:
83.81%
Total runs:
1.7K
Run Growth:
509
Growth Rate:
29.75%
Total runs:
1.7K
Run Growth:
-4.8K
Growth Rate:
-288.55%
Total runs:
1.6K
Run Growth:
466
Growth Rate:
28.64%
Total runs:
1.5K
Run Growth:
-1.4K
Growth Rate:
-93.45%
Total runs:
1.5K
Run Growth:
785
Growth Rate:
51.61%
Total runs:
1.5K
Run Growth:
1.5K
Growth Rate:
99.20%
Total runs:
1.5K
Run Growth:
-343
Growth Rate:
-23.16%
Total runs:
1.5K
Run Growth:
1.3K
Growth Rate:
89.69%
Total runs:
1.4K
Run Growth:
-1.4K
Growth Rate:
-95.23%
Total runs:
1.3K
Run Growth:
-1.5K
Growth Rate:
-118.32%
Total runs:
1.2K
Run Growth:
810
Growth Rate:
66.83%