Mungert/OLMo-2-0325-32B-Instruct-GGUF

OLMo-2-0325-32B-Instruct GGUF Models

Ultra-Low-Bit Quantization with IQ-DynamicGate (1-2 bit)

Our latest quantization method introduces precision-adaptive quantization for ultra-low-bit models (1-2 bit), with benchmark-proven improvements on Llama-3-8B. This approach uses layer-specific strategies to preserve accuracy while maintaining extreme memory efficiency.

Benchmark Context

All tests conducted on Llama-3-8B-Instruct using:

• Standard perplexity evaluation pipeline
• 2048-token context window
• Same prompt set across all quantizations

Method

• Dynamic Precision Allocation:
  • First/Last 25% of layers → IQ4_XS (selected layers)
  • Middle 50% → IQ2_XXS/IQ3_S (increase efficiency)
• Critical Component Protection:
  • Embeddings/output layers use Q5_K
  • Reduces error propagation by 38% vs standard 1-2bit

Quantization Performance Comparison (Llama-3-8B)

Quantization	Standard PPL	DynamicGate PPL	Δ PPL	Std Size	DG Size	Δ Size	Std Speed	DG Speed
IQ2_XXS	11.30	9.84	-12.9%	2.5G	2.6G	+0.1G	234s	246s
IQ2_XS	11.72	11.63	-0.8%	2.7G	2.8G	+0.1G	242s	246s
IQ2_S	14.31	9.02	-36.9%	2.7G	2.9G	+0.2G	238s	244s
IQ1_M	27.46	15.41	-43.9%	2.2G	2.5G	+0.3G	206s	212s
IQ1_S	53.07	32.00	-39.7%	2.1G	2.4G	+0.3G	184s	209s

Key:

• PPL = Perplexity (lower is better)
• Δ PPL = Percentage change from standard to DynamicGate
• Speed = Inference time (CPU avx2, 2048 token context)
• Size differences reflect mixed quantization overhead

Key Improvements:

• 🔥 IQ1_M shows massive 43.9% perplexity reduction (27.46 → 15.41)
• 🚀 IQ2_S cuts perplexity by 36.9% while adding only 0.2GB
• ⚡ IQ1_S maintains 39.7% better accuracy despite 1-bit quantization

Tradeoffs:

• All variants have modest size increases (0.1-0.3GB)
• Inference speeds remain comparable (<5% difference)

When to Use These Models

📌 Fitting models into GPU VRAM

✔ Memory-constrained deployments

✔ Cpu and Edge Devices where 1-2bit errors can be tolerated

✔ Research into ultra-low-bit quantization

Choosing the Right Model Format

Selecting the correct model format depends on your hardware capabilities and memory constraints.

BF16 (Brain Float 16) – Use if BF16 acceleration is available

• A 16-bit floating-point format designed for faster computation while retaining good precision.
• Provides similar dynamic range as FP32 but with lower memory usage.
• Recommended if your hardware supports BF16 acceleration (check your device's specs).
• Ideal for high-performance inference with reduced memory footprint compared to FP32.

📌 Use BF16 if:
✔ Your hardware has native BF16 support (e.g., newer GPUs, TPUs).
✔ You want higher precision while saving memory.
✔ You plan to requantize the model into another format.

📌 Avoid BF16 if:
❌ Your hardware does not support BF16 (it may fall back to FP32 and run slower).
❌ You need compatibility with older devices that lack BF16 optimization.

---

F16 (Float 16) – More widely supported than BF16

• A 16-bit floating-point high precision but with less of range of values than BF16.
• Works on most devices with FP16 acceleration support (including many GPUs and some CPUs).
• Slightly lower numerical precision than BF16 but generally sufficient for inference.

📌 Use F16 if:
✔ Your hardware supports FP16 but not BF16.
✔ You need a balance between speed, memory usage, and accuracy.
✔ You are running on a GPU or another device optimized for FP16 computations.

📌 Avoid F16 if:
❌ Your device lacks native FP16 support (it may run slower than expected).
❌ You have memory limitations.

---

Quantized Models (Q4_K, Q6_K, Q8, etc.) – For CPU & Low-VRAM Inference

Quantization reduces model size and memory usage while maintaining as much accuracy as possible.

• Lower-bit models (Q4_K) → Best for minimal memory usage, may have lower precision.
• Higher-bit models (Q6_K, Q8_0) → Better accuracy, requires more memory.

📌 Use Quantized Models if:
✔ You are running inference on a CPU and need an optimized model.
✔ Your device has low VRAM and cannot load full-precision models.
✔ You want to reduce memory footprint while keeping reasonable accuracy.

📌 Avoid Quantized Models if:
❌ You need maximum accuracy (full-precision models are better for this).
❌ Your hardware has enough VRAM for higher-precision formats (BF16/F16).

---

Very Low-Bit Quantization (IQ3_XS, IQ3_S, IQ3_M, Q4_K, Q4_0)

These models are optimized for extreme memory efficiency, making them ideal for low-power devices or large-scale deployments where memory is a critical constraint.

• IQ3_XS: Ultra-low-bit quantization (3-bit) with extreme memory efficiency.

  • Use case: Best for ultra-low-memory devices where even Q4_K is too large.
  • Trade-off: Lower accuracy compared to higher-bit quantizations.

• IQ3_S: Small block size for maximum memory efficiency.

  • Use case: Best for low-memory devices where IQ3_XS is too aggressive.

• IQ3_M: Medium block size for better accuracy than IQ3_S.

  • Use case: Suitable for low-memory devices where IQ3_S is too limiting.

• Q4_K: 4-bit quantization with block-wise optimization for better accuracy.

  • Use case: Best for low-memory devices where Q6_K is too large.

• Q4_0: Pure 4-bit quantization, optimized for ARM devices.

  • Use case: Best for ARM-based devices or low-memory environments.

---

Summary Table: Model Format Selection

Model Format	Precision	Memory Usage	Device Requirements	Best Use Case
BF16	Highest	High	BF16-supported GPU/CPUs	High-speed inference with reduced memory
F16	High	High	FP16-supported devices	GPU inference when BF16 isn't available
Q4_K	Medium Low	Low	CPU or Low-VRAM devices	Best for memory-constrained environments
Q6_K	Medium	Moderate	CPU with more memory	Better accuracy while still being quantized
Q8_0	High	Moderate	CPU or GPU with enough VRAM	Best accuracy among quantized models
IQ3_XS	Very Low	Very Low	Ultra-low-memory devices	Extreme memory efficiency and low accuracy
Q4_0	Low	Low	ARM or low-memory devices	llama.cpp can optimize for ARM devices

---

Included Files & Details

OLMo-2-0325-32B-Instruct-bf16.gguf

• Model weights preserved in BF16.
• Use this if you want to requantize the model into a different format.
• Best if your device supports BF16 acceleration.

OLMo-2-0325-32B-Instruct-f16.gguf

• Model weights stored in F16.
• Use if your device supports FP16, especially if BF16 is not available.

OLMo-2-0325-32B-Instruct-bf16-q8_0.gguf

• Output & embeddings remain in BF16.
• All other layers quantized to Q8_0.
• Use if your device supports BF16 and you want a quantized version.

OLMo-2-0325-32B-Instruct-f16-q8_0.gguf

• Output & embeddings remain in F16.
• All other layers quantized to Q8_0.

OLMo-2-0325-32B-Instruct-q4_k.gguf

• Output & embeddings quantized to Q8_0.
• All other layers quantized to Q4_K.
• Good for CPU inference with limited memory.

OLMo-2-0325-32B-Instruct-q4_k_s.gguf

• Smallest Q4_K variant, using less memory at the cost of accuracy.
• Best for very low-memory setups.

OLMo-2-0325-32B-Instruct-q6_k.gguf

• Output & embeddings quantized to Q8_0.
• All other layers quantized to Q6_K .

OLMo-2-0325-32B-Instruct-q8_0.gguf

• Fully Q8 quantized model for better accuracy.
• Requires more memory but offers higher precision.

OLMo-2-0325-32B-Instruct-iq3_xs.gguf

• IQ3_XS quantization, optimized for extreme memory efficiency.
• Best for ultra-low-memory devices.

OLMo-2-0325-32B-Instruct-iq3_m.gguf

• IQ3_M quantization, offering a medium block size for better accuracy.
• Suitable for low-memory devices.

OLMo-2-0325-32B-Instruct-q4_0.gguf

• Pure Q4_0 quantization, optimized for ARM devices.
• Best for low-memory environments.
• Prefer IQ4_NL for better accuracy.

🚀 If you find these models useful

❤ Please click "Like" if you find this useful!
Help me test my AI-Powered Network Monitor Assistant with quantum-ready security checks:
👉 Free Network Monitor

💬 How to test:

1. Click the chat icon (bottom right on any page)
2. Choose an AI assistant type:
   • TurboLLM (GPT-4-mini)
   • FreeLLM (Open-source)
   • TestLLM (Experimental CPU-only)

What I’m Testing

I’m pushing the limits of small open-source models for AI network monitoring, specifically:

• Function calling against live network services
• How small can a model go while still handling:
  • Automated Nmap scans
  • Quantum-readiness checks
  • Metasploit integration

🟡 TestLLM – Current experimental model (llama.cpp on 6 CPU threads):

• ✅ Zero-configuration setup
• ⏳ 30s load time (slow inference but no API costs)
• 🔧 Help wanted! If you’re into edge-device AI, let’s collaborate!

Other Assistants

🟢 TurboLLM – Uses gpt-4-mini for:

• Real-time network diagnostics
• Automated penetration testing (Nmap/Metasploit)
• 🔑 Get more tokens by downloading our Free Network Monitor Agent

🔵 HugLLM – Open-source models (≈8B params):

• 2x more tokens than TurboLLM
• AI-powered log analysis
• 🌐 Runs on Hugging Face Inference API

💡 Example AI Commands to Test:

1. "Give me info on my websites SSL certificate"
2. "Check if my server is using quantum safe encyption for communication"
3. "Run a quick Nmap vulnerability test"

OLMo 2 32B Instruct March 2025 is post-trained variant of the OLMo-2 32B March 2025 model, which has undergone supervised finetuning on an OLMo-specific variant of the Tülu 3 dataset, further DPO training on this dataset, and final RLVR training on this 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 or Tülu 3 paper for more details!

OLMo is a series of Open Language Models designed to enable the science of language models. These models are trained on the Dolma dataset. We are releasing all code, checkpoints, logs, and associated training details.

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/OLMo-2-0325-32B-DPO

Model Sources

• Project Page: https://allenai.org/olmo
• Repositories:
  • Core repo (training, inference, fine-tuning etc.): https://github.com/allenai/OLMo-core
  • Evaluation code: https://github.com/allenai/olmes
  • Further fine-tuning code: https://github.com/allenai/open-instruct
• Paper: https://arxiv.org/abs/2501.00656
• Demo: https://playground.allenai.org/

Installation

OLMo 2 will be supported in the next version of Transformers, and you need to install it from the main branch using:

pip install --upgrade git+https://github.com/huggingface/transformers.git

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-2-0325-32B-Instruct")

Chat template

NOTE: This is different than previous OLMo 2 and Tülu 3 models due to a minor change in configuration. It does NOT have the bos token before the rest. Our other models have <|endoftext|> at the beginning of the chat template.

The chat template for our models is formatted as:

<|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:

<|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.

Intermediate Checkpoints

To facilitate research on RL finetuning, we have released our intermediate checkpoints during the model's RLVR training. The model weights are saved every 20 training steps, and can be accessible in the revisions of the HuggingFace repository. For example, you can load with:

olmo_model = AutoModelForCausalLM.from_pretrained("allenai/OLMo-2-0325-32B-Instruct", revision="step_200")

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 2 LC	BBH	DROP	GSM8k	IFEval	MATH	MMLU	Safety	PopQA	TruthQA
Closed API models
GPT-3.5 Turbo 0125	59.6	38.7	66.6	70.2	74.3	66.9	41.2	70.2	69.1	45.0	62.9
GPT 4o Mini 2024-07-18	65.7	49.7	65.9	36.3	83.0	83.5	67.9	82.2	84.9	39.0	64.8
Open weights models
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
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
Gemma-2-27b-it	61.3	49.0	72.7	67.5	80.7	63.2	35.1	70.7	75.9	33.9	64.6
Qwen2.5-32B	66.5	39.1	82.3	48.3	87.5	82.4	77.9	84.7	82.4	26.1	70.6
Mistral-Small-24B	67.6	43.2	80.1	78.5	87.2	77.3	65.9	83.7	66.5	24.4	68.1
Llama-3.1-70B	70.0	32.9	83.0	77.0	94.5	88.0	56.2	85.2	76.4	46.5	66.8
Llama-3.3-70B	73.0	36.5	85.8	78.0	93.6	90.8	71.8	85.9	70.4	48.2	66.1
Gemma-3-27b-it	-	63.4	83.7	69.2	91.1	-	-	81.8	-	30.9	-
Fully open models
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
OLMo-2-32B-0325-SFT	61.7	16.9	69.7	77.2	78.4	72.4	35.9	76.1	93.8	35.4	61.3
OLMo-2-32B-0325-DPO	68.8	44.1	70.2	77.5	85.7	83.8	46.8	78.0	91.9	36.4	73.5
OLMo-2-32B-0325-Instruct	68.8	42.8	70.6	78.0	87.6	85.6	49.7	77.3	85.9	37.5	73.2

Learning curves

Below is the training curves for allenai/OLMo-2-0325-32B-Instruct. The model was trained using 5 8xH100 nodes.

Below are the core eval scores over steps for allenai/OLMo-2-0325-32B-Instruct (note we took step 320 as the final checkpoint, corresponding to episode 573,440):

Below are the other eval scores over steps for allenai/OLMo-2-0325-32B-Instruct:

Reproduction command

The command below is copied directly from the tracked training job:

# clone and check out commit
git clone https://github.com/allenai/open-instruct.git
# this should be the correct commit, the main thing is to have the vllm monkey patch for
# 32b olmo https://github.com/allenai/open-instruct/blob/894ffa236319bc6c26c346240a7e4ee04ba0bd31/open_instruct/vllm_utils2.py#L37-L59
git checkout a51dc98525eec01de6e8a24c071f42dce407d738
uv sync
uv sync --extra compile

# note that you may need 5 8xH100 nodes for the training.
# so please setup ray properly, e.g., https://github.com/allenai/open-instruct/blob/main/docs/tulu3.md#llama-31-tulu-3-70b-reproduction
python open_instruct/grpo_vllm_thread_ray_gtrl.py \
    --exp_name 0310_olmo2_32b_grpo_12818 \
    --beta 0.01 \
    --local_mini_batch_size 32 \
    --number_samples_per_prompt 16 \
    --output_dir output \
    --local_rollout_batch_size 4 \
    --kl_estimator kl3 \
    --learning_rate 5e-7 \
    --dataset_mixer_list allenai/RLVR-GSM-MATH-IF-Mixed-Constraints 1.0 \
    --dataset_mixer_list_splits train \
    --dataset_mixer_eval_list allenai/RLVR-GSM-MATH-IF-Mixed-Constraints 16 \
    --dataset_mixer_eval_list_splits train \
    --max_token_length 2048 \
    --max_prompt_token_length 2048 \
    --response_length 2048 \
    --model_name_or_path allenai/OLMo-2-0325-32B-DPO \
    --non_stop_penalty \
    --stop_token eos \
    --temperature 1.0 \
    --ground_truths_key ground_truth \
    --chat_template_name tulu \
    --sft_messages_key messages \
    --eval_max_length 4096 \
    --total_episodes 10000000 \
    --penalty_reward_value 0.0 \
    --deepspeed_stage 3 \
    --no_gather_whole_model \
    --per_device_train_batch_size 2 \
    --local_rollout_forward_batch_size 2 \
    --actor_num_gpus_per_node 8 8 8 4 \
    --num_epochs 1 \
    --vllm_tensor_parallel_size 1 \
    --vllm_num_engines 12 \
    --lr_scheduler_type constant \
    --apply_verifiable_reward true \
    --seed 1 \
    --num_evals 30 \
    --save_freq 20 \
    --reward_model_multiplier 0.0 \
    --no_try_launch_beaker_eval_jobs \
    --try_launch_beaker_eval_jobs_on_weka \
    --gradient_checkpointing \
    --with_tracking

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. This model has been fine-tuned using a dataset mix with outputs generated from third party models and are subject to additional terms: Gemma Terms of Use.

Citation

@article{olmo20242olmo2furious,
      title={2 OLMo 2 Furious}, 
      author={Team OLMo and Pete Walsh and Luca Soldaini and Dirk Groeneveld and Kyle Lo and Shane Arora and Akshita Bhagia and Yuling Gu and Shengyi Huang and Matt Jordan and Nathan Lambert and Dustin Schwenk and Oyvind Tafjord and Taira Anderson and David Atkinson and Faeze Brahman and Christopher Clark and Pradeep Dasigi and Nouha Dziri and Michal Guerquin and Hamish Ivison and Pang Wei Koh and Jiacheng Liu and Saumya Malik and William Merrill and Lester James V. Miranda and Jacob Morrison and Tyler Murray and Crystal Nam and Valentina Pyatkin and Aman Rangapur and Michael Schmitz and Sam Skjonsberg and David Wadden and Christopher Wilhelm and Michael Wilson and Luke Zettlemoyer and Ali Farhadi and Noah A. Smith and Hannaneh Hajishirzi},
      year={2024},
      eprint={2501.00656},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2501.00656}, 
}