Standalone merged BF16 model from epoch 3.970178926441352, step 1500. Dataset: CompassioninMachineLearning/urban12738cleaned at ef7c0e742df63ea319e35d02d9f6ba63d6e7c68d. Training: 10,072 distinct documents plus 2,000 repeat exposures per epoch; 200 disjoint validation documents. Merged with Unsloth's native savepretrainedmerged(savemethod="merged16bit"). Weights are validated BF16 and packaged losslessly into eight safetensors shards. No adapter is required to load this model. See runmanifest.json for base revision, document selection hashes, training parameters and export validation. Training does not establish an improvement in compassion; evaluate that separately.
Open-weight model · Text generation
OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08
by Neel Rajani Neelectric/OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08
OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08 is an open-weight model for text generation from Neel Rajani. It has 7.3B parameters and a 4,096-token context. At 16-bit it needs about 17.5 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
This model is a fine-tuned version of allenai/OLMo-2-1124-7B-Instruct. It has been trained using TRL. This model was trained with SFT.
Parameters7.3B
Context4,096
Weights14.6 GB
License—
AccessOpen weights
Monthly Downloads—
Runs On
What it takes to serve OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08 (7.3B 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 | 14.6 GB | 17.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 7.3 GB | 8.8 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 3.6 GB | 4.4 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 Oct 23, 2026.
Model Card
This model is a fine-tuned version of allenai/OLMo-2-1124-7B-Instruct. It has been trained using TRL. This model was trained with SFT.
Excerpt from the card by Neel Rajani.
Configuration
- Architecture
- Olmo2ForCausalLM
- Context length (tokens)
- 4,096
- Layers
- 32
- Hidden size
- 4,096
- Feed-forward size
- 11,008
- Attention heads
- 32
- Key/value heads
- 32
- Vocabulary size
- 100,352
- RoPE base
- 500,000
- Model type
- olmo2
Identity and Version
- Repository
- Neelectric/OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08
- Publisher
- Neel Rajani
- Task
- Text generation
- Modality
- Text
- Library
- transformers
- Parameters
- 7.3B parameters
- Languages
- sft, trl
- Revision
- d0b2cd997dce65da2d3ea710200d404416c7ca04
- First published
- 2026-10-08
- Last updated
- 2026-10-09
Files and Weights
15 files, 14.6 GB in total. The weights are 4 files totalling 14.6 GB in bin, safetensors.
Weights4 files · 14.6 GB
Configuration4 files · 30.9 KB
Tokenizer4 files · 9.7 MB
Documentation1 file · 1.7 KB
Other1 file · 1.1 KB
Repository1 file · 1.5 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00003.safetensors | Weights | 5.0 GB | 1604007a9637 |
| model-00002-of-00003.safetensors | Weights | 5.0 GB | 0690638bdac7 |
| model-00003-of-00003.safetensors | Weights | 4.6 GB | e938ada6fef0 |
| training_args.bin | Weights | 9.4 KB | 16e3cd923c7d |
| config.json | Configuration | 646 B | — |
| generation_config.json | Configuration | 157 B | — |
| model.safetensors.index.json | Configuration | 29.6 KB | — |
| special_tokens_map.json | Configuration | 467 B | — |
| README.md | Documentation | 1.7 KB | — |
| chat_template.jinja | Other | 1.1 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| merges.txt | Tokenizer | 916.6 KB | — |
| tokenizer.json | Tokenizer | 7.1 MB | — |
| tokenizer_config.json | Tokenizer | 4.3 KB | — |
| vocab.json | Tokenizer | 1.6 MB | — |
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 14.6 GB
Download from Neel Rajani
Released by Neel Rajani through its official repository on Hugging Face.
Built From
- Derived from allenai/OLMo-2-1124-7B-Instruct
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 14.6 GB |
| 16-bit | 14.6 GB |
| 8-bit | 7.3 GB |
| 4-bit | 3.6 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08
How much GPU memory does OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08 need?
About 17.5 GB at 16-bit and 4.4 GB at 4-bit: the weights (7.3B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08 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 OLMo-2-1124-7B-Instruct_SFT_mathsp_ewc_v00.08's context length?
4,096 tokens, from the maximum position embeddings in its published configuration.
Similar Models
An experimental post-training transfer study. We applied the Vinci SFT + DPO character recipe to mistralai/Mistral-7B-Instruct-v0.3 to answer one question: does character training developed on a different model lineage transfer to this one? Apache-2.0, 7.25B, drop-in with transformers. This uses a retired base, and we are saying so first. Mistral lists Mistral 7B Instruct v0.3 as retired as of 30 March 2025 (deprecated 30 November 2024), with Ministral 3 8B as the recommended replacement. "Retired" is Mistral's own lifecycle term. The open weights remain downloadable on Hugging Face under Apache-2.0. We selected this base for continuity with our earlier experiments, not because it is…
A 4-bit quantized, business-focused conversational AI built on Mistral 7B using Unsloth. Optimized for low VRAM (runs on 6GB+ VRAM) and for Indian business use-cases. - 7B params in 4-bit (BF16 base): ~4GB RAM me chal jayega from unsloth import FastLanguageModel model, tokenizer = FastLanguageModel.frompretrained( modelname = "UX4567/Khushi-Business-AI-4bit", maxseqlength = 2048, loadin4bit = True, prompt = "Ek customer bol raha hai: Mujhe refund chahiye, kya reply du?" inputs = tokenizer([prompt], returntensors = "pt").to("cuda") outputs = model.generate(inputs, maxnewtokens = 200) print(tokenizer.batchdecode(outputs)[0])
Korean instruction-following language model. Built from Yejin 8B v3 (pretrained from scratch by Yejin: cooler8/yejin-korean-8b-v3-base): continued pretraining for 6,000 steps on public Korean data, Yejin chat vector from our own SFT model, then LoRA (rank 64, lr 2e-4). The LoRA data mixes our phase-1 format data (45,013 samples) with 39,290 short Korean solutions written by Qwen/Qwen3.6-27B (Apache-2.0) for questions from our own puzzle generator, KMMLU train, AI Hub school exams and AI Hub reading comprehension; only solutions that reached the known answer were kept (sequence-level distillation; no teacher weights or logits). No benchmark test items were used, and the teacher solutions…
The Mistral-7B-Instruct-v0.2 Large Language Model (LLM) is an instruct fine-tuned version of the Mistral-7B-v0.2. Mistral-7B-v0.2 has the following changes compared to Mistral-7B-v0.1 - 32k context window (vs 8k context in v0.1) - Rope-theta = 1e6 For full details of this model please read our paper and release blog post. In order to leverage instruction fine-tuning, your prompt should be surrounded by [INST] and [/INST] tokens. The very first instruction should begin with a begin of sentence id. The next instructions should not. The assistant generation will be ended by the end-of-sentence token id. This format is available as a chat template via the applychattemplate() method: - If you…
A Mistral-7B-Instruct-v0.2 checkpoint compressed with SVD-LLM to 70.0% of dense parameters, then edited by 10 of 10 rounds of iterative parameter-neutral swap selected by the gapiter rule (up to 0.1% of dense parameters per round; the full run's budget is 1.0%). This is a research artifact from a study of how SVD compression damages safety behaviour and which component-selection rule best repairs it. It is one cell of a grid over selection rules and budgets; it is not a general-purpose chat model. This checkpoint exists to measure safety/utility trade-offs under compression. Several arms in the grid are deliberately safety-degraded relative to Mistral-7B-Instruct-v0.2: compression alone…