# Emendator by Aidan: Open-Weight Model Source: https://savrn.com/models/emendator Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Emendator (3.7B 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 | 7.5 GB | 9.0 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 3.7 GB | 4.5 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 1.9 GB | 2.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/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](https://savrn.com/ai-index/pricing/gpus), read Oct 7, 2026. [Emendator on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/emendator/gpus) ## Model Card Emendator is a byt5-xl model finetuned to correct OCR artifacts in Latin text. This model cannot provide completely faithful reconstruction for all orthographies - on a large scale, it will shift the distribution of tokens towards that which it has been trained on. This is to say: Emendator will take editorial liberties with your data. As such, use it only in circumstances when the primary concern is only to recover intelligible Latin, not to recover intelligible Latin of a particular style. The model is intended to be used on segments of 250 characters. Anything else will compromise performance. Original: "atque optimo viro, peterem; superavi tamen dignitate Catilinam, gratia Galbam. Quod… Excerpt from the card by Aidan, licensed cc-by-nc-4.0. ## Configuration Architecture T5ForConditionalGeneration Vocabulary size 384 Stored precision float32 Model type t5 ## Identity and Version Repository aimgo/Emendator Publisher Aidan Task Text generation Modality Text Library Not stated by the source Parameters 3.7B parameters Languages la Revision 52040eff93121a80d63e3ee25a56714f578a21d2 First published 2026-01-02 Last updated 2026-09-27 ## Files and Weights 16 files, 15.0 GB in total. The weights are 7 files totalling 15.0 GB in bin, pt, pth, safetensors. Weights7 files · 15.0 GB Configuration6 files · 85.2 KB Tokenizer1 file · 25.6 KB Documentation1 file · 3.6 KB Repository1 file · 1.5 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-00003.safetensors | Weights | 5.0 GB | c317a30f5538 | | model-00002-of-00003.safetensors | Weights | 5.0 GB | 08832c6299f1 | | model-00003-of-00003.safetensors | Weights | 5.0 GB | de54e2d9a158 | | optimizer.pt | Weights | 11.3 MB | ebb5eb45d073 | | rng_state.pth | Weights | 14.2 KB | e3bd9aeb359c | | scheduler.pt | Weights | 1.1 KB | 14152aafb58c | | training_args.bin | Weights | 7.7 KB | 70df7de9410a | | added_tokens.json | Configuration | 3.0 KB | — | | config.json | Configuration | 823 B | — | | generation_config.json | Configuration | 207 B | — | | model.safetensors.index.json | Configuration | 45.1 KB | — | | special_tokens_map.json | Configuration | 3.1 KB | — | | trainer_state.json | Configuration | 33.0 KB | — | | README.md | Documentation | 3.6 KB | — | | .gitattributes | Repository | 1.5 KB | — | | tokenizer_config.json | Tokenizer | 25.6 KB | — | ## License and Download License cc-by-nc-4.0 Access Open weights, no gate Download size 15.0 GB [Download from Aidan](https://huggingface.co/aimgo/Emendator) Released by Aidan through its official repository on Hugging Face. [Read the license](https://creativecommons.org/licenses/by-nc/4.0/). ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 15.0 GB | | 16-bit | 7.5 GB | | 8-bit | 3.7 GB | | 4-bit | 1.9 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Questions About Emendator ### How much GPU memory does Emendator need? About 9 GB at 16-bit and 2.2 GB at 4-bit: the weights (3.7B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Emendator 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 Emendator commercially? Not without separate permission. Emendator is released under Creative Commons Attribution-NonCommercial 4.0. CC BY-NC 4.0 permits sharing and adapting with credit for non-commercial purposes only. Commercial use needs separate permission from the rights holder. ## Similar Models Model · Text generation ### [granite-4.2-3b-MXFP4A16-GPTQ](https://savrn.com/models/granite-4-2-3b-mxfp4a16-gptq) [David Reedy](https://savrn.com/model-publishers/quark75) An experimental MXFP4 quantization of ibm-granite/granite-4.2-3b, made with GPTQ calibration instead of the usual data-free rounding, and provided in two formats with identical weights. It's published as a data point, not as the recommended build. For most users, the Q4KM GGUF or GPTQ int4 build is the better choice. Format: OCP MX — E2M1 (FP4) weights, groupsize=32, one UE8M0 shared scale per group; weights only, activations stay 16-bit. Embeddings, norms and lmhead are unquantized. MXFP4 is usually produced data-free, by rounding weights straight to FP4 with no calibration data. llm-compressor ships an MXFP4A16 scheme, but every example we found pairs GPTQModifier with integer schemes… Open weights apache-2.0 3.7B parameters 131,072 tokens transformers [View model](https://savrn.com/models/granite-4-2-3b-mxfp4a16-gptq) Model · Text generation ### [granite-4.2-3b-W4A16-GPTQ-gs32](https://savrn.com/models/granite-4-2-3b-w4a16-gptq-gs32) [David Reedy](https://savrn.com/model-publishers/quark75) A 4-bit weight-only GPTQ quantization of ibm-granite/granite-4.2-3b, saved in compressed-tensors format for vLLM. On an AMD Radeon Pro V620 (RDNA2), vLLM served it at 65 tok/s single-sequence and ~1,000 tok/s aggregate at 16 concurrent sequences. llm-compressor GPTQModifier: - int4, symmetric, weight-only (activations stay 16-bit) - groupsize=32, actorder="weight" - lmhead left unquantized 512 samples, maxseqlength=512 The tuning mattered. With the default W4A16 preset (groupsize=128), the same pipeline scored 0.153 mean KLD. Moving to groupsize=32 with actorder="weight" cut that by about 28%, for roughly 0.2 GB more disk. For comparison, an AWQ build (W4A16 asymmetric, groupsize=128) from… Open weights apache-2.0 3.7B parameters 131,072 tokens transformers [View model](https://savrn.com/models/granite-4-2-3b-w4a16-gptq-gs32) Model · Text generation ### [phi4-clinical-mlx](https://savrn.com/models/phi4-clinical-mlx) [Web](https://savrn.com/model-publishers/charakaweb) A specialized 3.8B biomedical & clinical reasoning model built on Microsoft's Phi-4-mini-instruct, optimized natively for Apple Silicon Metal acceleration via Apple MLX. The model underwent a 3-stage transfer learning curriculum: 1. Stage 1 (STEM Foundation): 116,000 instruction pairs across NCERT Classes 6–12 (Physics, Chemistry, Biology) eliminating foundational science hallucinations. 2. Stage 2 (PubMed 2026 Evidence): 12 recent 2026 clinical update archives from NCBI FTP covering survival outcomes (OS, PFS, HR), targeted therapeutics, and clinical trial endpoints. 3. Stage 3 (Comprehensive Internal Medicine): Balanced multi-specialty clinical curriculum (cardiology, nephrology… Open weights mit 3.8B parameters 131,072 tokens mlx [View model](https://savrn.com/models/phi4-clinical-mlx) Model · Text generation ### [phi4-clinical](https://savrn.com/models/phi4-clinical) [Web](https://savrn.com/model-publishers/charakaweb) A specialized 3.8B biomedical & clinical reasoning foundation model built on Microsoft's Phi-4-mini-instruct, formatted for standard Hugging Face transformers and PyTorch. 1. Stage 1 (STEM Foundation): 116,000 instruction pairs across NCERT Classes 6–12 (Physics, Chemistry, Biology) eliminating foundational science hallucinations. 2. Stage 2 (PubMed 2026 Evidence): 12 recent 2026 clinical update archives from NCBI FTP covering survival outcomes (OS, PFS, HR), targeted therapeutics, and clinical trial endpoints. 3. Stage 3 (Comprehensive Internal Medicine): Balanced multi-specialty clinical curriculum (cardiology, nephrology, endocrinology, pulmonology) with an active oncology replay buffer.… Open weights mit 3.8B parameters 131,072 tokens transformers [View model](https://savrn.com/models/phi4-clinical) Model · Text generation ### [StandardOne-3B](https://savrn.com/models/standardone-3b) [Standard Thinking](https://savrn.com/model-publishers/standardthinking) Standard One scores a bounded set of answers for a supplied scenario and returns probabilities through POST /v1/systemone. It does not generate free-form response text. This repository contains the merged BF16 3B checkpoint; the server code is in StandardOne-8B. In the reported served evaluations, 3B has a lower median latency on the measured short-request profile; 8B scores higher on the public standard and hard tiers. See Benchmarks for the measurement conditions and limitations. The figure combines results from different measurement paths. See Benchmarks for served versus offline conditions; measured 24–26 September 2026. - Send a state and a bounded rubric to receive probabilities for… Open weights apache-2.0 3.8B parameters 262,144 tokens transformers [View model](https://savrn.com/models/standardone-3b) Model · Text generation ### [StandardOne-3B-FP8](https://savrn.com/models/standardone-3b-fp8) [Standard Thinking](https://savrn.com/model-publishers/standardthinking) StandardOne-3B-FP8 is an FP8 (compressed-tensors, float8e4m3 weights, dynamic per-token activations) quantization of the released StandardOne-3B decision model. The language-model linear projections (q/k/v/o, gate/up/down) are quantized per-channel FP8 E4M3 with dynamic FP8 activations (llm-compressor's data-free FP8DYNAMIC recipe, no calibration data required); the vision tower, multi-modal projector, embeddings and lmhead are left unquantized in BF16. It was produced from source revision 68dafd17ead9b8cf6f85c4f08f7f2f3a1e7b9e5c of StandardOne-3B on 2026-09-25 using llm-compressor 0.14.0 (torch 2.14.0, transformers 5.17.0, compressed-tensors 0.19.0); results below. Served through SGLang… Open weights apache-2.0 3.8B parameters 262,144 tokens transformers [View model](https://savrn.com/models/standardone-3b-fp8) ## Aidan [All models and datasets](https://savrn.com/model-publishers/aimgo) ## Versions - [52040eff9312](https://savrn.com/models/emendator/versions/52040eff9312) · current 2026-09-27 ## Explore More - [All text generation models](https://savrn.com/models/tasks/text-generation) - [All models under cc-by-nc-4.0](https://savrn.com/models/licenses/cc-by-nc-4-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-09-27. - [Hugging Face record](https://huggingface.co/aimgo/Emendator) - [How the hub is built](https://savrn.com/model-hub/methodology)