Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Built upon extensive training, Qwen3 delivers groundbreaking advancements in reasoning, instruction-following, agent capabilities, and multilingual support, with the following key features: - Uniquely support of seamless switching between thinking mode (for complex logical reasoning, math, and coding) and non-thinking mode (for efficient, general-purpose dialogue) within single model, ensuring optimal performance across various scenarios. - Significantly enhancement in its reasoning capabilities, surpassing previous QwQ (in thinking mode) and…
Open-weight model · Text generation
dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918
by Minjae Oh Riasok/dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918
This model is a fine-tuned version of cosmos1030/gmp-kd3e-1-s80pct-lr1e-420260916220740 on the trl-lib/ultrafeedbackbinarized dataset. It has been trained using TRL.
Parameters4B
Context40,960
Weights8.0 GB
License—
AccessOpen weights
Monthly Downloads—
Runs On
What it takes to serve dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918 (4B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 8.0 GB | 9.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 4.0 GB | 4.8 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 2.0 GB | 2.4 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
Model Card
This model is a fine-tuned version of cosmos1030/gmp-kd3e-1-s80pct-lr1e-420260916220740 on the trl-lib/ultrafeedbackbinarized dataset. It has been trained using TRL. This model was trained with DPO, a method introduced in Direct Preference Optimization: Your Language Model is Secretly a Reward Model.
Excerpt from the card by Minjae Oh.
Configuration
- Architecture
- Qwen3ForCausalLM
- Context length (tokens)
- 40,960
- Layers
- 36
- Hidden size
- 2,560
- Feed-forward size
- 9,728
- Attention heads
- 32
- Key/value heads
- 8
- Head dimension
- 128
- Vocabulary size
- 151,936
- RoPE base
- 1,000,000
- Stored precision
- bfloat16
- Model type
- qwen3
Identity and Version
- Repository
- Riasok/dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918
- Publisher
- Minjae Oh
- Task
- Text generation
- Modality
- Text
- Library
- transformers
- Parameters
- 4B parameters
- Languages
- Not stated by the source
- Revision
- de83e0509ff4a7d2bf4a970095ec725c4a7e0a10
- First published
- 2026-09-18
- Last updated
- 2026-09-18
Files and Weights
18 files, 8.1 GB in total. The weights are 2 files totalling 8.0 GB in safetensors.
Weights2 files · 8.0 GB
Configuration9 files · 1.0 MB
Tokenizer4 files · 15.9 MB
Documentation1 file · 2.8 KB
Other1 file · 4.2 KB
Repository1 file · 1.6 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00002.safetensors | Weights | 5.0 GB | 5fb5d1de7f03 |
| model-00002-of-00002.safetensors | Weights | 3.1 GB | c1095c34327b |
| added_tokens.json | Configuration | 707 B | — |
| all_results.json | Configuration | 190 B | — |
| config.json | Configuration | 1.5 KB | — |
| eval_summary_resumed.json | Configuration | 530 B | — |
| generation_config.json | Configuration | 214 B | — |
| model.safetensors.index.json | Configuration | 32.9 KB | — |
| special_tokens_map.json | Configuration | 613 B | — |
| train_results.json | Configuration | 190 B | — |
| trainer_state.json | Configuration | 983.8 KB | — |
| README.md | Documentation | 2.8 KB | — |
| chat_template.jinja | Other | 4.2 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| merges.txt | Tokenizer | 1.7 MB | — |
| tokenizer.json | Tokenizer | 11.4 MB | aeb13307a71a |
| tokenizer_config.json | Tokenizer | 5.4 KB | — |
| vocab.json | Tokenizer | 2.8 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 8.0 GB
Download from Minjae Oh
Released by Minjae Oh through its official repository on Hugging Face.
Built From
- Derived from cosmos1030/gmp-kd3e-1-s80pct-lr1e-4_20260916_220740
- Described by arXiv:2305.18290
- Trained on (disclosed) trl-lib/ultrafeedback_binarized
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 8.0 GB |
| 16-bit | 8.0 GB |
| 8-bit | 4.0 GB |
| 4-bit | 2.0 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918
How much GPU memory does dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918 need?
About 9.7 GB at 16-bit and 2.4 GB at 4-bit: the weights (4B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What is dpo-scout-s80-ultrafeedback-lr1e-5-beta0.05-epoch1.0_20260918's context length?
40,960 tokens, from the maximum position embeddings in its published configuration.
Similar Models
We introduce the updated version of the Qwen3-4B non-thinking mode, named Qwen3-4B-Instruct-2507, featuring the following key enhancements: - Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage. - Substantial gains in long-tail knowledge coverage across multiple languages. - Markedly better alignment with user preferences in subjective and open-ended tasks, enabling more helpful responses and higher-quality text generation. - Enhanced capabilities in 256K long-context understanding. Qwen3-4B-Instruct-2507 has the following features: NOTE: This model supports only non-thinking…
Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Building upon extensive advancements in training data, model architecture, and optimization techniques, Qwen3 delivers the following key improvements over the previously released Qwen2.5: Qwen3-4B-Base has the following features: For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our blog, GitHub, and Documentation. The code of Qwen3 has been in the latest Hugging Face transformers and we advise you to use the latest version of transformers. With transformers<4.51.0, you will…
Qwen3 is the latest generation of large language models in Qwen series, offering a comprehensive suite of dense and mixture-of-experts (MoE) models. Building upon extensive advancements in training data, model architecture, and optimization techniques, Qwen3 delivers the following key improvements over the previously released Qwen2.5: Qwen3-4B-Base has the following features: For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our blog, GitHub, and Documentation. The code of Qwen3 has been in the latest Hugging Face transformers and we advise you to use the latest version of transformers. With transformers<4.51.0, you will…
We introduce the updated version of the Qwen3-4B non-thinking mode, named Qwen3-4B-Instruct-2507, featuring the following key enhancements: - Significant improvements in general capabilities, including instruction following, logical reasoning, text comprehension, mathematics, science, coding and tool usage. - Substantial gains in long-tail knowledge coverage across multiple languages. - Markedly better alignment with user preferences in subjective and open-ended tasks, enabling more helpful responses and higher-quality text generation. - Enhanced capabilities in 256K long-context understanding. Qwen3-4B-Instruct-2507 has the following features: NOTE: This model supports only non-thinking…
The full ImmuneCoT method fuses the two safety branches with a Base-adjusted product-of-experts Qimm(v) ∝ q̃rec(v)·q̃res(v)/qB(v). This checkpoint uses the naive fusion Qno-base(v) ∝ q̃rec(v)·q̃res(v) — the same branch weights (λrec=0.5, λres=0.7) but no division by the base distribution — isolating whether the gains come from combining Recognition+Response at all, or specifically from the Base-adjusted PoE term. Intended use: research reproducibility for the ImmuneCoT paper's RQ3 ablation.