Text Generation Models

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.18.1

SNIF (StateGate Neural Inference Fabric) is an open-source composite causal language model that dynamically connects Qwen2.5-0.5B (Edge Prefill and Fast Decode) and DeepSeek-R1-1.5B (Cognitive Reasoning) via a continuous 2048-dimensional latent interconnect. Note: Research paper and extended theoretical evaluation in preparation.

Model · Text generation

spark-darwin-17b-3way

Charlie

Darwin uniform-soup of 3 Spark2.5-1.7B donors (226/226 tensors verified). Agentic/tool-use + uncensored + multilingual blend. Distillation donor for the 4B program. Needs transformers==5.17.0 + trustremotecode=True.

Model · Text generation

sql-qwen3.5-9b-v9-obfuscated-1.0-match-std-2epochs-output-sql

Mariia Ponomarenko

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). - PEFT 0.21.0

Model · Text generation

ST1.0

Sterling Ai

Hopefully the Next best SAFE agentic AI for long coding tasks. With the edition of Being better at code, less hallucination, less lies, less Confidently lies. I want to make this for the scope of a AI that offloads the coding tasks so AI can do the hard coding tasks wile the humans do stuff humans do best. As the saying something along the lines of "AI should be doing my chores wile i make art instead of me doing chores wile AI makes art" anyways I am trying to make this Model card more professional also if you want to give feedback before this moddel enteres training go on https://sterlingai.wordpress.com to awnser on how you want ai to improve. here's some planned ideas for what it mite…

Model · Text generation

stormdesk-triage

Nathan DeMoss

This qwen3 model was trained 2x faster with Unsloth and Huggingface's TRL library.

This repository contains the final epoch-4 LoRA adapter from the 1.7B all-reference STRIDE experiment. The base model is Qwen/Qwen3-1.7B, used with thinking disabled during training and evaluation. Each prompt produced eight rollouts. Novelty was computed against eligible steps from all eight rollouts, including correct and incorrect trajectories. Only correct trajectories received the auxiliary STRIDE diversity bonus; incorrect trajectories retained their ordinary outcome/GRPO signal. This adapter uses local-positive step allocation. Use the Qwen chat template with enablethinking=False to reproduce the non-thinking evaluation setting. This is a research adapter for mathematical reasoning.…

STRIDE local step credit: all eligible rollouts define novelty; only correct rollouts receive the bonus. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The…

GRPO with group-standardized answer correctness, without a STRIDE diversity bonus. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE ablation using one pooled reasoning embedding per eligible response. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is 64…

STRIDE ablation distributing each response diversity bonus uniformly over its eligible tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global…

STRIDE local step credit: all eligible rollouts define novelty; only correct rollouts receive the bonus. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The…

STRIDE token-uniform: all eligible rollouts define novelty; only correct rollouts receive a mass-preserving uniform token bonus. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the…

GRPO with group-standardized answer correctness, without a STRIDE diversity bonus. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE with nonnegative local step-diversity credit on eligible reasoning tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is…

STRIDE ablation using one pooled reasoning embedding per eligible response. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global prompt batch is 64…

STRIDE ablation distributing each response diversity bonus uniformly over its eligible tokens. Training explicitly renders enablethinking=False, recorded as thinkingmode: false in the scientific contract. The pinned tokenizer and chat template remain unchanged. Use the same explicit keyword at inference, particularly for Qwen3-1.7B, whose default template enables thinking. Every checkpoint in this repository requires explicit nonthinking provenance. This public repository retains every published optimizer-update adapter, including update zero (the initial, untrained adapter). Training is planned for 4 epochs on the same 2,048-question training split as the earlier STRIDE runs. The global…

SUPERBRAIN-1 is a 9B Arabic model by ABUZ8, fine-tuned from QCRI/Fanar-1-9B to do Arabic root, conjugate, derive masdars and participles, and name a verb's form. It is a completion model, not a chat model: give it one of the prompt formats below and read the first line of the output. Packet B — held-out test. 1,343 items over 258 roots that were Quranic Arabic Corpus) and modern words from the Arramooz lexicon. Root items are asked both with a few-shot prompt and as a bare question. Root accuracy by root class (bare): sound 79%, hollow 60%, doubled 56%, hamzated 51%, defective 45%, assimilated 44%. 85-item gate (v3) — 50 root words, 20 arithmetic/logic questions, 15 Generalization probe.…

A Llama-2-7b-chat checkpoint compressed with SVD-LLM to 60.0% of dense parameters, then given a 0.0% parameter budget of restored SVD components selected by the unknown rule. 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 rate, and the point of the study is to quantify that and test recovery. Treat…

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…

A Llama-3-8B-Instruct 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 60.0% of dense parameters, then edited by 8 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 60.0% of dense parameters, then edited by 9 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 1 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 2 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 3 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 4 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 5 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 6 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 7 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% of dense parameters, then edited by 8 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-3-8B-Instruct: compression alone raises…

A Llama-3-8B-Instruct checkpoint compressed with SVD-LLM to 50.0% 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-3-8B-Instruct: compression alone raises…

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…

This model is a fine-tuned version of francesca9805/ppt-wc-loglinear-newlex-swa-before-100mb-packed-bfdisoseed3407. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn100mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn10mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn10mb. It has been trained using TRL. This model was trained with SFT.

This model is a fine-tuned version of goldfish-models/swalatn10mb. It has been trained using TRL. This model was trained with SFT.

Official GGUF quantized weights for swift-1.5-27b-coding, a specialized 27B parameter coding model fine-tuned using DoRA on bjivanovich/code-py-rust-cpp-50k (covering Python, Rust, C++, and multi-step reasoning). All quantizations were generated using an importance matrix (imatrix) calibrated directly on domain-specific programming samples.

Model · Text generation

Swift-1.5-Qwen3.8-27b-heretic-MXFP4-DFlash2-radiance

SlyBase

A single Radiance container (.rad, 18.05 GiB) holding the target model and its speculative drafter, quantised for AMD RDNA4 (Radeon AI PRO R9700, gfx1201, 32 GB). It is not a Transformers/vLLM checkpoint: it runs only with the Radiance inference engine. The GPTQ source checkpoint is published separately as 1. Qwen/Qwen3.8-27B (Apache-2.0) 2. ukisai/Swift-1.5-Qwen3.8-27b (Swift Open License v1.0) 3. akumaburn/Swift-1.5-Qwen3.8-27b-heretic (abliterated, BF16) 4. GPTQ to MXFP4 (llm-compressor, MXFP4A16, static act-order, 89 calibration samples, lmhead/vision/MTP excluded — gptq-recipe.yaml) 5. rad-convert (Radiance 1.2.3) with the DFlash2 drafter incoai/Qwen3.8-27B-DFlash2 (Apache-2.0) merged…

Model · Text generation

Swift-1.5-Qwen3.8-27B-MXFP4-GPTQ

SlyBase

quantised to OCP MXFP4 (fp4e2m1 weights, e8m0 scales, group size 32) with GPTQ, in the compressed-tensors mxfp4-pack-quantized format. Weight-only: activations stay in 16-bit. This is the source checkpoint of the Radiance container and the checkpoint the vllm-sly-radiance image is benchmarked with (together with the DFlash2 drafter syvai/Qwen3.8-27B-DFlash2-W4A16). 1. Qwen/Qwen3.8-27B (Apache-2.0) 2. ukisai/Swift-1.5-Qwen3.8-27b (Swift Open License v1.0) 3. akumaburn/Swift-1.5-Qwen3.8-27b-heretic (abliterated, BF16) 4. This checkpoint: llm-compressor GPTQModifier, scheme MXFP4A16, static act-order, sequential pipeline, 89 calibration samples, max sequence length 2048 (recipe.yaml) The…

Model · Text generation

Swift-1.5-Qwen3.8-27B-MXFP4-radiance

SlyBase

A single Radiance container (.rad, 16.79 GiB) with the target model only, quantised for AMD RDNA4 (Radeon AI PRO R9700, gfx1201, 32 GB). It is not a Transformers/vLLM checkpoint. It is the drafter-free sibling of which carries the DFlash2 drafter in the same file; use this one with the model's own MTP block, with a different drafter, or without speculative decoding. The GPTQ source checkpoint is 1. Qwen/Qwen3.8-27B (Apache-2.0) 2. ukisai/Swift-1.5-Qwen3.8-27b (Swift Open License v1.0) 3. akumaburn/Swift-1.5-Qwen3.8-27b-heretic (abliterated, BF16) 4. GPTQ to MXFP4 (llm-compressor, MXFP4A16, static act-order, 89 calibration samples, lmhead/vision/MTP excluded — gptq-recipe.yaml) 5.…

An uncensored community variant of Swift 1.5 Qwen3.8-27B in native NInfer format, with mixed NVFP4/FP8 weights and embedded DFlash2 speculative decoding. This is a third-party release, not an official UkisAI or OrcaRouter checkpoint. SHA-256: d094126721091481d8fa3b34f780eb68fcaac61ef0750b698cb5076b6b2e84b1 Use a compatible NInfer build that supports Qwen3.8-27B NVFP4 and DFlash2. See the upstream project for installation and supported hardware. This file is not GGUF and cannot be loaded directly with Transformers or llama.cpp. Add --vision for image input. Increase --max-context as available VRAM allows; the example starts at 8,192 tokens. DFlash2 is already embedded, so no separate draft…

An uncensored community variant of Swift Qwen3.8-27B in native NInfer format, with mixed NVFP4/FP8 weights and embedded DFlash2 speculative decoding. This is a third-party release, not an official UkisAI or OrcaRouter checkpoint. SHA-256: b57bb82f6c43fa4bba3bcb883f45ccff12b81fd461502a37c79d78c3ba7a979b Use a compatible NInfer build that supports Qwen3.8-27B NVFP4 and DFlash2. See the upstream project for installation and supported hardware. This file is not GGUF and cannot be loaded directly with Transformers or llama.cpp. Add --vision for image input. Increase --max-context as available VRAM allows; the example starts at 8,192 tokens. DFlash2 is already embedded, so no separate draft…