Open-weight model · Question answering
tyness-boe-1.2b-pre-alpha-3-instruct
by Jason Jiménez Cruz jasonjimnz/tyness-boe-1.2b-pre-alpha-3-instruct
tyness-boe-1.2b-pre-alpha-3-instruct is an open-weight model for question answering from Jason Jiménez Cruz, released under other. Its published files total 6.7 GB.
Third pre-alpha from Tyness Train, an agent-guided fine-tuning harness for small LLMs on low-VRAM GPUs.
Parameters—
Context—
Weights6.7 GB
Licenseother
AccessOpen weights
Monthly Downloads—
Model Card
Third pre-alpha from Tyness Train, an agent-guided fine-tuning harness for small LLMs on low-VRAM GPUs. This model is RAG-style grounded (answers only from provided norm excerpts) and adds two confirmed recipe upgrades: a structured METADATOS block in the system prompt (official identifiers, dates, rank, department, article-title index) and adversarial correction pairs (user asserts a false premise → the model corrects it with the exact citation). Evaluated on fresh, uncontaminated teacher-generated sets: 200 questions per context style (never part of any training split), each grounded in a specific precept, judged by an LLM referee (glm5.3-flash): correct / partial / incorrect + 0–1 score.…
Excerpt from the card by Jason Jiménez Cruz, licensed other.
Identity and Version
- Repository
- jasonjimnz/tyness-boe-1.2b-pre-alpha-3-instruct
- Publisher
- Jason Jiménez Cruz
- Task
- Question answering
- Modality
- Text
- Library
- peft
- Parameters
- Not stated by the source
- Languages
- es
- Revision
- 4eb17c22cd0b14c300c8d28e3075d31d82ae4d38
- First published
- 2026-10-12
- Last updated
- 2026-10-12
Files and Weights
45 files, 6.7 GB in total. The weights are 16 files totalling 6.7 GB in bin, gguf, pt, pth, safetensors.
Weights16 files · 6.7 GB
Configuration10 files · 37.0 KB
Tokenizer8 files · 18.9 MB
Documentation6 files · 25.1 KB
Other4 files · 21.9 KB
Repository1 file · 1.7 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| adapter/adapter_model.safetensors | Weights | 1.8 MB | 0f1a5d344c46 |
| adapter/training_args.bin | Weights | 5.8 KB | 9bea9872196f |
| checkpoints/checkpoint-328/adapter_model.safetensors | Weights | 1.8 MB | 7f5607137566 |
| checkpoints/checkpoint-328/optimizer.pt | Weights | 7.1 MB | 1a68c02714a1 |
| checkpoints/checkpoint-328/rng_state.pth | Weights | 14.6 KB | ce28501a5280 |
| checkpoints/checkpoint-328/scheduler.pt | Weights | 1.5 KB | 2a8820b630ab |
| checkpoints/checkpoint-328/training_args.bin | Weights | 5.8 KB | 9bea9872196f |
| checkpoints/checkpoint-410/adapter_model.safetensors | Weights | 1.8 MB | 0f1a5d344c46 |
| checkpoints/checkpoint-410/optimizer.pt | Weights | 7.1 MB | 56975099978b |
| checkpoints/checkpoint-410/rng_state.pth | Weights | 14.6 KB | 42ec52ae8e1c |
| checkpoints/checkpoint-410/scheduler.pt | Weights | 1.5 KB | 171d7ffd721c |
| checkpoints/checkpoint-410/training_args.bin | Weights | 5.8 KB | 9bea9872196f |
| gguf/merged-bf16-f16.gguf | Weights | 2.3 GB | a190f98b2fcc |
| gguf/merged-bf16-q4_k_m.gguf | Weights | 730.9 MB | 94f9b6a765f9 |
| gguf/merged-bf16-q8_0.gguf | Weights | 1.2 GB | db3c9289266c |
| merged-bf16/model.safetensors | Weights | 2.3 GB | 71a74a8ad9c7 |
| adapter/adapter_config.json | Configuration | 1.1 KB | — |
| adapter/run_plan.json | Configuration | 572 B | — |
| checkpoints/checkpoint-328/adapter_config.json | Configuration | 1.1 KB | — |
| checkpoints/checkpoint-328/trainer_state.json | Configuration | 10.2 KB | — |
| checkpoints/checkpoint-410/adapter_config.json | Configuration | 1.1 KB | — |
| checkpoints/checkpoint-410/trainer_state.json | Configuration | 12.8 KB | — |
| hf_manifest.json | Configuration | 7.6 KB | — |
| merged-bf16/config.json | Configuration | 1.3 KB | — |
| merged-bf16/generation_config.json | Configuration | 132 B | — |
| run_config.json | Configuration | 1.0 KB | — |
| README.md | Documentation | 7.3 KB | — |
| adapter/README.md | Documentation | 5.2 KB | — |
| checkpoints/README.md | Documentation | 1.4 KB | — |
| checkpoints/checkpoint-328/README.md | Documentation | 5.2 KB | — |
| checkpoints/checkpoint-410/README.md | Documentation | 5.2 KB | — |
| eval-report.md | Documentation | 751 B | — |
| adapter/chat_template.jinja | Other | 5.5 KB | — |
| checkpoints/checkpoint-328/chat_template.jinja | Other | 5.5 KB | — |
| checkpoints/checkpoint-410/chat_template.jinja | Other | 5.5 KB | — |
| merged-bf16/chat_template.jinja | Other | 5.5 KB | — |
| .gitattributes | Repository | 1.7 KB | — |
| adapter/tokenizer.json | Tokenizer | 4.7 MB | — |
| adapter/tokenizer_config.json | Tokenizer | 519 B | — |
| checkpoints/checkpoint-328/tokenizer.json | Tokenizer | 4.7 MB | — |
| checkpoints/checkpoint-328/tokenizer_config.json | Tokenizer | 519 B | — |
| checkpoints/checkpoint-410/tokenizer.json | Tokenizer | 4.7 MB | — |
| checkpoints/checkpoint-410/tokenizer_config.json | Tokenizer | 519 B | — |
| merged-bf16/tokenizer.json | Tokenizer | 4.7 MB | — |
| merged-bf16/tokenizer_config.json | Tokenizer | 519 B | — |
License and Download
- License
- other
- Access
- Open weights, no gate
- Download size
- 6.7 GB
Download from Jason Jiménez Cruz
Released by Jason Jiménez Cruz through its official repository on Hugging Face. Read the license.
Built From
- Adapter of LiquidAI/LFM2.5-1.2B-Instruct
- Derived from LiquidAI/LFM2.5-1.2B-Instruct
- Trained on (disclosed) boe-focused-adv-v1
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 6.7 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About tyness-boe-1.2b-pre-alpha-3-instruct
What license is tyness-boe-1.2b-pre-alpha-3-instruct released under?
other, as its publisher declares it. Read the license text before commercial use.
Similar Models
Splinter-base is the pretrained model discussed in the paper Few-Shot Question Answering by Pretraining Span Selection (at ACL 2021). Its original repository can be found here. The model is case-sensitive. Note: This model doesn't contain the pretrained weights for the QASS layer (see paper for details), and therefore the QASS layer is randomly initialized upon loading it. For the model with those weights, see tau/splinter-base-qass. Splinter is a model that is pretrained in a self-supervised fashion for few-shot question answering. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an…
This model can be used for the task of question answering. The model should not be used to intentionally create hostile or alienating environments for people. Significant research has explored bias and fairness issues with language models (see, e.g., Sheng et al. (2021) and Bender et al. (2021)). Predictions generated by the model may include disturbing and harmful stereotypes across protected classes; identity characteristics; and sensitive, social, and occupational groups. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. The model creators note in the associated paper: The model creators note in the associated paper: The model…
This model is a fine-tune checkpoint of DistilBERT-base-cased, fine-tuned using (a second step of) knowledge distillation on SQuAD v1.1. This model reaches a F1 score of 87.1 on the dev set (for comparison, BERT bert-base-cased version reaches a F1 score of 88.7).
This model can be used for the task of Question Answering on Legal Documents. Read: An Open Source Contractual Language Understanding Application Using Machine Learning for detailed information on training procedure, dataset preprocessing and evaluation. See CUAD dataset card for more information. See CUAD dataset card for more information. Used V100/P100 from Google Colab Pro Python, Transformers Mohammed Rakib in collaboration with Ezi Ozoani and the Hugging Face team Use the code below to get started with the model.