Community abliteration and quantization of tsinghua-sigs-robot-lab/VeriLoop-E2, fixed revision 9379d199adcc44acdf836558037163878e3a37ae. This is not an official Tsinghua release. Model architecture is Qwen35ForConditionalGeneration; the upstream names the base Qwen3.8-27B. AutoRound W4A16, symmetric group 128, 256 public Pile calibration samples, 512 tokens, 200 iterations, batch 4. Eligible weights are INT4; small hybrid projections, activations and retained weights are BF16. All source tensors are represented; the 333 vision tensors are preserved exactly. A bounded six-candidate Heretic 1.4.0 search selected per-layer strength 1.2; 96 of 128 targeted residual-output matrices changed and…
Open-weight model · Text generation
Hemmingway-1-AutoRound-W4G128
by Facu Vlad J groxaxo/Hemmingway-1-AutoRound-W4G128
Hemmingway-1-AutoRound-W4G128 is an open-weight model for text generation from Facu Vlad J, released under Apache License 2.0. It has 5.8B parameters and a 262,144-token context. At 16-bit it needs about 14 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index.
4-bit, group-size-128 AutoRound quantization of Altworld/Hemmingway-1, a 27B-parameter fine-tune of Qwen3.8-27B. This repository is an independent quantization, not the original model release.
Parameters5.8B
Context262,144
Weights18.6 GB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads—
Runs On
What it takes to serve Hemmingway-1-AutoRound-W4G128 (5.8B 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 | 11.6 GB | 14.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 5.8 GB | 7.0 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 2.9 GB | 3.5 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 22, 2026.
Hemmingway-1-AutoRound-W4G128 on every accelerator the SAVRN Index prices, at every precision
Model Card
By Facu Vlad J, published under apache-2.0, revision 9e7a457c81b8.
4-bit, group-size-128 AutoRound quantization of Altworld/Hemmingway-1, a 27B-parameter fine-tune of Qwen3.8-27B. This repository is an independent quantization, not the original model release. See the original model card for training, intended use, and benchmark information. Quantization: AutoRound 0.15.1, bits=4, groupsize=128, packingformat=autoround:autoawq, seqlen=1024, nsamples=64, iters=50. Certain layers remain in higher precision, as specified in quantizationconfig.json. The included six-shard weight index and auxiliary modelextratensors.safetensors are needed together. Approximate download size: 18.65 GB (17.37 GiB). This is a quantization of an existing fine-tuned model, not a new…
Read Facu Vlad J's full model card
4-bit, group-size-128 AutoRound quantization of Altworld/Hemmingway-1, a 27B-parameter fine-tune of Qwen3.8-27B. This repository is an independent quantization, not the original model release. See the original model card for training, intended use, and benchmark information.
Quantization: AutoRound 0.15.1, bits=4, group_size=128, packing_format=auto_round:auto_awq, seqlen=1024, nsamples=64, iters=50. Certain layers remain in higher precision, as specified in quantization_config.json.
The included six-shard weight index and auxiliary model_extra_tensors.safetensors are needed together. Approximate download size: 18.65 GB (17.37 GiB). This is a quantization of an existing fine-tuned model, not a new fine-tuning run. Inference compatibility depends on a Transformers/AutoRound stack supporting this model architecture and quantization format; no inference benchmark or quality claim is made here.
Configuration
- Architecture
- Qwen3_5ForCausalLM
- Context length (tokens)
- 262,144
- Layers
- 64
- Hidden size
- 5,120
- Feed-forward size
- 17,408
- Attention heads
- 24
- Key/value heads
- 4
- Head dimension
- 256
- Vocabulary size
- 248,320
- Stored precision
- float16
- Model type
- qwen3_5_text
- Quantization
- auto-round
Identity and Version
- Repository
- groxaxo/Hemmingway-1-AutoRound-W4G128
- Publisher
- Facu Vlad J
- Task
- Text generation
- Modality
- Text
- Library
- transformers
- Parameters
- 5.8B parameters
- Languages
- awq
- Revision
- 9e7a457c81b8ba453c4285d21df5f2832543f264
- First published
- 2026-09-24
- Last updated
- 2026-09-24
Files and Weights
16 files, 18.6 GB in total. The weights are 7 files totalling 18.6 GB in safetensors.
Weights7 files · 18.6 GB
Configuration4 files · 164.7 KB
Tokenizer2 files · 20.0 MB
Documentation1 file · 1.2 KB
Other1 file · 9.1 KB
Repository1 file · 1.6 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| model-00001-of-00006.safetensors | Weights | 3.2 GB | 88efbe3ccb07 |
| model-00002-of-00006.safetensors | Weights | 3.2 GB | 31dec76a13c8 |
| model-00003-of-00006.safetensors | Weights | 3.2 GB | c625602805f4 |
| model-00004-of-00006.safetensors | Weights | 3.1 GB | 27e1f8f4589a |
| model-00005-of-00006.safetensors | Weights | 2.5 GB | 3e0c9e8dbe1d |
| model-00006-of-00006.safetensors | Weights | 2.5 GB | 6866cf8adccc |
| model_extra_tensors.safetensors | Weights | 849.4 MB | 1d8268aa85ac |
| config.json | Configuration | 13.7 KB | — |
| generation_config.json | Configuration | 227 B | — |
| model.safetensors.index.json | Configuration | 140.8 KB | — |
| quantization_config.json | Configuration | 10.0 KB | — |
| README.md | Documentation | 1.2 KB | — |
| chat_template.jinja | Other | 9.1 KB | — |
| .gitattributes | Repository | 1.6 KB | — |
| tokenizer.json | Tokenizer | 20.0 MB | 06b9509352d2 |
| tokenizer_config.json | Tokenizer | 1.2 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 18.6 GB
Download from Facu Vlad J
Released by Facu Vlad J through its official repository on Hugging Face. Read the license.
Built From
- Derived from Altworld/Hemmingway-1
- Quantized from Altworld/Hemmingway-1
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 18.6 GB |
| 16-bit | 11.6 GB |
| 8-bit | 5.8 GB |
| 4-bit | 2.9 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Hemmingway-1-AutoRound-W4G128
How much GPU memory does Hemmingway-1-AutoRound-W4G128 need?
About 14 GB at 16-bit and 3.5 GB at 4-bit: the weights (5.8B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Hemmingway-1-AutoRound-W4G128 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use Hemmingway-1-AutoRound-W4G128 commercially?
Yes. Hemmingway-1-AutoRound-W4G128 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
What is Hemmingway-1-AutoRound-W4G128's context length?
262,144 tokens, from the maximum position embeddings in its published configuration.
Similar Models
Qwen3-Coder is available in multiple sizes. Today, we're excited to introduce Qwen3-Coder-30B-A3B-Instruct. This streamlined model maintains impressive performance and efficiency, featuring the following key enhancements: - Significant Performance among open models on Agentic Coding, Agentic Browser-Use, and other foundational coding tasks. - Long-context Capabilities with native support for 256K tokens, extendable up to 1M tokens using Yarn, optimized for repository-scale understanding. - Agentic Coding supporting for most platform such as Qwen Code, CLINE, featuring a specially designed function call format. Qwen3-Coder-30B-A3B-Instruct has the following features: NOTE: This model…
Vigyan-1.5B 4× MoE is an experimental Sparse Mixture of Experts model upcycled from 4 specialized domain LoRA adapters (Science, Technology, Engineering, Mathematics). It was assembled to evaluate whether post-hoc expert stitching on compact language models ($\le 3\text{B}$) can deliver multi-domain specialization at single-expert inference latency with a sub-1GB RAM footprint. A pre-quantized standalone GGUF (vigyan-1.5b-4x-moe.Q4KM.gguf) is available directly within this repository for instantaneous local execution on CPU or mobile. Test this model on Google Colab with an embedded Gradio chat interface: - shreyansh12183/vigyan-1.5b-adapter-science…
The NVIDIA Qwen2.5-VL-7B-Instruct-FP4 model is the quantized version of Alibaba's Qwen2.5-VL-7B-Instruct model, which is an auto-regressive language model that uses an optimized transformer architecture. For more information, please check here. The NVIDIA Qwen2.5-VL-7B-Instruct-FP4 model is quantized with TensorRT Model Optimizer. This model is ready for commercial/non-commercial use. This model is not owned or developed by NVIDIA. It was developed and built to a third party’s requirements for this application and use case. See the Non-NVIDIA (Qwen2.5-VL-7B-Instruct) Model Card. Use of this model is governed by nvidia-open-model-license Global, except in European Union Developers looking to…
Chirp Chirp! We are introducing Ornith-1.5, a major step toward building foundation models through end-to-end self-improvement. Ornith-1.5 extends Ornith-1.0 (which was developed on top of Qwen3.5 and Gemma4 with additional continued pretraining, mid-training, and post-training) by expanding the self-improvement loop from scaffold and rollout optimization to jointly optimizing task generation, scaffold construction, and solution rollouts. Rather than relying on a fixed set of human-curated tasks and manually designed harnesses, Ornith-1.5 continuously generates new training tasks, discovers effective strategies for solving them, and improves the policy through reinforcement learning. For…
A Llama-2-7b-chat checkpoint compressed with SVD-LLM to 60.0% 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 Llama-2-7b-chat: compression alone raises attack-success…