README: lead with reasoning effort, state what ships, link quants, say benchmarks are unrun

README.md

@@ -1,241 +1,266 @@
----
-license: apache-2.0
-language:
-  - en
-pipeline_tag: image-text-to-text
-base_model: Qwen/Qwen3.8-27B
-library_name: transformers
-tags:
-  - multimodal
-  - vision-language
-  - reasoning
-  - thinking
-  - efficient-reasoning
-  - code
-  - software-engineering
-  - swe
-  - agentic
-  - terminal
-  - tool-use
-  - long-context
-  - qwen3.8
-  - thinking-efficiency
-model-index:
-  - name: Salience-27B-R5
-    results: []
----
-
-<div align="center">
-
-# Salience — 27B
-<p align="center">
-  <img src="Salience-27B-R5-Banner.png" alt="Vection Labs Salience 27B R5 Banner" width="100%">
-</p>
-
-**A 27B dense vision-language engineer that stops thinking once it has the answer.**
-
-*Vection Labs*
-
-[Weights](https://huggingface.co/vectionlabs/Salience-27B-R5) ·
-[Reasoning effort](#reasoning-effort) ·
-[Quickstart](#quickstart) ·
-[Limitations](#limitations--responsible-use)
-
-</div>
-
----
-> [!Note]
-> **R5.** Fifth revision of the Salience Ridge 27B tier, rebuilt on the Qwen3.8 architecture.
-> Stable for daily use; rough edges get fixed in the stable release — report them in the
-> Community tab.
-
-## Abstract
-
-Salience 27B is a **27-billion-parameter dense** vision-language model built for **hard,
-practical engineering work**: writing and debugging real code, repo-scale edits, multi-step
-terminal agency, and quantitative reasoning — with native vision and **1,048,576 tokens** of
-context.
-
-Where the MoE tiers of the family (Pro, Flash) route a few billion active parameters per token,
-Salience 27B runs **all 27B on every token** — maximum per-token capacity, a hybrid
-linear+full attention stack for long-context speed, and an **MTP head** for self-speculative
-decoding.
-
-R5's headline change is **reasoning economy**. A reasoning model pays for accuracy in tokens,
-and most of them pay the same price for *"what does this flag do"* as for *"why does this
-deadlock under load"*. R5 does not: it reasons hard when the problem needs it and answers
-directly when it does not — and unlike the stock configuration, that is the **default**
-behaviour rather than something you have to ask for.
-
-## Highlights
-
-- **Reasoning economy by default.** Deliberation is proportional to difficulty. The model is no
-  longer instructed to validate assumptions and weigh alternatives on every single turn — it
-  decides. Ask for depth explicitly and you still get it.
-- **Dense capacity.** All 27B parameters active on every token — no routing, no expert misses,
-  maximum depth on every step of a hard problem.
-- **SWE-agent first.** Tuned to produce runnable code, repo-scale edits, methodical debugging,
-  and well-formed native tool calls.
-- **Lives in a terminal.** Plans the command sequence, checks each result before the next step,
-  and recovers from failures instead of repeating them.
-- **A million tokens.** Paste the repository, not the fragment.
-- **Genuinely multimodal.** Images and video are first-class inputs — read a diagram, a UI
-  screenshot, a stack-trace screenshot, or a whiteboard photo mid-task.
-- **Fast decode for its size.** Hybrid linear+full attention (full every 4th layer) plus an MTP
-  head for self-speculative decoding.
-- **Direct.** Reduced refusal behaviour: it answers the question you asked. See
-  [responsible use](#limitations--responsible-use).
-- **Open weights.** Apache-2.0, `transformers`-native.
-
-## Model overview
-
-| | |
-|---|---|
-| **Parameters** | 27.8B dense (all active) |
-| **Modalities** | text, image, video -> text |
-| **Context window** | 1,048,576 tokens (YaRN + Dual Chunk Attention) |
-| **Attention** | hybrid linear + full attention (full every 4th layer) |
-| **Decoding** | MTP head included (self-speculative decoding) |
-| **Precision** | bfloat16 |
-| **Architecture** | Qwen3.8 dense (27B) + native vision encoder |
-| **License** | Apache-2.0 |
-| **Library** | 🤗 `transformers` (`AutoModelForImageTextToText`) |
-
-The family: [Pro (35B-A3B MoE)](https://huggingface.co/vectionlabs/Salience-1.5-Pro) ·
-[Flash (30B-A3B MoE)](https://huggingface.co/vectionlabs/Salience-1.5-Flash) ·
-**27B R5 (dense)** ·
-[Nano (9B dense)](https://huggingface.co/vectionlabs/Salience-1.5-Nano)
-
-## Capabilities
-
-- **Code & SWE execution** — runnable code, repo-scale edits, methodical debugging, robust backends.
-- **Terminal & agentic work** — multi-step planning, tool orchestration, long-horizon task execution.
-- **Deep reasoning** — structured, inspectable chains of thought for hard, multi-step problems.
-- **Multimodal perception** — diagrams, screenshots, documents, and video as first-class inputs.
-
-## Reasoning effort
-
-Thinking is **on by default**: the model reasons inside `<think>...</think>` before answering,
-and serving stacks expose it as `reasoning_content`. What R5 changes is **how much**.
-
-| value | behaviour | use it for |
-|---|---|---|
-| `low` | keeps the chain short and moves straight to the conclusion | chat, lookups, formatting, refactors |
-| `medium` | **default** — no deliberation instruction; the model decides | everyday engineering work |
-| `xhigh` | deliberate at length, validate assumptions, weigh alternatives | hard debugging, architecture, math |
-
-```python
-# default: proportional reasoning, nothing to configure
-text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-
-# ask for depth when the problem earns it
-text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True,
-                                reasoning_effort="xhigh")
-
-# skip thinking entirely
-text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True,
-                                enable_thinking=False)
-```
-
-Reasoning is native — you never have to write *think step by step*. Doing so makes a model of
-this kind *perform* reasoning instead of doing it.
-
-## Tool calling
-
-The model emits **XML-style tool calls** (`<tool_call><function=...><parameter=...>`), parsed
-natively by vLLM / SGLang tool parsers for this model family, and by `llama-server --jinja`.
-Provide tool schemas via the chat template `tools` argument.
-
-## Intended use
-
-Salience 27B R5 targets **software engineering, coding agents, and technical research**:
-
-- Code generation, explanation, debugging, review, and repo-scale tasks.
-- Terminal / tool-using agent workflows (CLI agents, browsing, ML engineering, DevOps).
-- Backend and systems design, infrastructure-as-code.
-- Step-by-step reasoning and quantitative problem solving.
-- Screenshot / diagram / document understanding inside engineering workflows.
-
-It is **not** intended for high-stakes decisions without human review, nor as a source of truth
-for medical, legal, or financial advice.
-
-## Quickstart
-
-```python
-from transformers import AutoModelForImageTextToText, AutoProcessor
-import torch
-
-repo = "vectionlabs/Salience-27B-R5"
-proc = AutoProcessor.from_pretrained(repo)
-model = AutoModelForImageTextToText.from_pretrained(
-    repo, dtype="auto", device_map="auto"
-)
-
-messages = [{
-    "role": "user",
-    "content": [{"type": "text", "text": "Implement an LRU cache in Python with O(1) get/put."}],
-}]
-text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-inputs = proc(text=[text], return_tensors="pt").to(model.device)
-out = model.generate(**inputs, max_new_tokens=2048)
-print(proc.batch_decode(out[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)[0])
-```
-
-Requires a recent `transformers` (>= 5.8). Vision works the same way with
-`{"type": "image", "image": ...}` content items.
-
-## Quantized GGUF (local)
-
-```bash
-llama-server -m Salience-27B-R5-Q4_K_M.gguf \
-  --jinja --reasoning-format deepseek \
-  -c 32768 -ngl 999
-```
-
-`--jinja` is not optional for agent use: it applies the model's own chat template, which is
-what turns XML tool calls into proper OpenAI-style `tool_calls` — and what makes the reasoning
-defaults above take effect. Without it you get malformed calls and stock behaviour.
-
-This is a **dense** model, so standard quant intuition applies: **Q4_K_M and up** hold quality
-well; use Q5_K_M/Q6_K when VRAM allows. (The MoE tiers of the family need Q5/Q6 minimum — that
-constraint does *not* apply here.) Keep the MTP layers if your quant includes them: they enable
-self-speculative decoding for free extra speed.
-
-## Long context
-
-Ships with YaRN (`factor 4.0`, `original_max_position_embeddings 262144`) and a
-`dual_chunk_attention_config` block. Static YaRN taxes short prompts slightly; that is the cost
-of having the full window available by default. vLLM and SGLang read the DCA block,
-`transformers` ignores it.
-
-## Prompting tips
-
-- **Let it think.** No "think step by step" — reasoning is native. Reach for `reasoning_effort`
-  instead of prompt scaffolding.
-- **Give it the repo.** A million tokens: paste whole files or repositories, not fragments.
-- **Agentic loops.** Use `--jinja` with llama-server (or vLLM/SGLang parsers) so XML tool calls
-  become proper OpenAI-style `tool_calls`.
-- **Vision mid-task.** Screenshots of stack traces and UI states work as debugging inputs.
-
-## Benchmarks
-
-Published when they come from a run that reproduces. Not before.
-
-## Limitations & responsible use
-
-- May hallucinate APIs or facts under ambiguity; verify critical output.
-- Review generated code before running it, especially anything touching production systems.
-- **Reduced refusal behaviour.** There is no content filter in the weights and no system-level
-  guardrail — the model will attempt requests a stock model declines, and it will not decline
-  on your behalf. Whatever policy your deployment needs is yours to add at the application
-  layer. You are responsible for what you generate and for complying with the law where you
-  operate.
-- `medium` reasoning by default means shorter chains on genuinely hard problems than a model
-  pinned to maximum effort. Pass `reasoning_effort="xhigh"` when the problem deserves it.
-
----
-
-<sub>Built on Qwen3.8 (Apache-2.0).</sub>
-<sub>Build with love by the vectionlabs' team (Apache-2.0).</sub>
-
-<div align="center"><sub>© 2026 Vection Labs</sub></div>
+---
+license: apache-2.0
+language:
+  - en
+pipeline_tag: image-text-to-text
+base_model: Qwen/Qwen3.8-27B
+library_name: transformers
+tags:
+  - multimodal
+  - vision-language
+  - reasoning
+  - thinking
+  - efficient-reasoning
+  - code
+  - software-engineering
+  - swe
+  - agentic
+  - terminal
+  - tool-use
+  - long-context
+  - qwen3.8
+  - thinking-efficiency
+model-index:
+  - name: Salience-27B-R5
+    results: []
+---
+
+<div align="center">
+
+# Salience — 27B
+<p align="center">
+  <img src="Salience-27B-R5-Banner.png" alt="Vection Labs Salience 27B R5 Banner" width="100%">
+</p>
+
+**A 27B dense vision-language engineer that stops thinking once it has the answer.**
+
+*Vection Labs*
+
+[Weights](https://huggingface.co/vectionlabs/Salience-27B-R5) ·
+[Reasoning effort](#reasoning-effort) ·
+[Quickstart](#quickstart) ·
+[Limitations](#limitations--responsible-use)
+
+</div>
+
+---
+> [!Note]
+> **R5.** Fifth revision of the Salience Ridge 27B tier, rebuilt on the Qwen3.8 architecture.
+> Stable for daily use; rough edges get fixed in the stable release — report them in the
+> Community tab.
+
+## Abstract
+
+Salience 27B is a **27-billion-parameter dense** vision-language model built for **hard,
+practical engineering work**: writing and debugging real code, repo-scale edits, multi-step
+terminal agency, and quantitative reasoning — with native vision and **1,048,576 tokens** of
+context.
+
+Where the MoE tiers of the family (Pro, Flash) route a few billion active parameters per token,
+Salience 27B runs **all 27B on every token** — maximum per-token capacity, a hybrid
+linear+full attention stack for long-context speed, and an **MTP head** for self-speculative
+decoding.
+
+R5's headline change is **reasoning economy**. A reasoning model pays for accuracy in tokens,
+and most of them pay the same price for *"what does this flag do"* as for *"why does this
+deadlock under load"*. R5 does not: it reasons hard when the problem needs it and answers
+directly when it does not — and unlike the stock configuration, that is the **default**
+behaviour rather than something you have to ask for.
+
+## Reasoning effort
+
+Thinking is **on by default**: the model reasons inside `<think>...</think>` before answering,
+and serving stacks expose it as `reasoning_content`. What R5 changes is **how much**.
+
+| value | behaviour | use it for |
+|---|---|---|
+| `low` | keeps the chain short and moves straight to the conclusion | chat, lookups, formatting, refactors |
+| `medium` | **default** — no deliberation instruction; the model decides | everyday engineering work |
+| `xhigh` | deliberate at length, validate assumptions, weigh alternatives | hard debugging, architecture, math |
+
+```python
+# default: proportional reasoning, nothing to configure
+text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+
+# ask for depth when the problem earns it
+text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True,
+                                reasoning_effort="xhigh")
+
+# skip thinking entirely
+text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True,
+                                enable_thinking=False)
+```
+
+Reasoning is native — you never have to write *think step by step*. Doing so makes a model of
+this kind *perform* reasoning instead of doing it.
+
+## Highlights
+
+- **Reasoning economy by default.** Deliberation is proportional to difficulty. The model is no
+  longer instructed to validate assumptions and weigh alternatives on every single turn — it
+  decides. Ask for depth explicitly and you still get it.
+- **Dense capacity.** All 27B parameters active on every token — no routing, no expert misses,
+  maximum depth on every step of a hard problem.
+- **SWE-agent first.** Tuned to produce runnable code, repo-scale edits, methodical debugging,
+  and well-formed native tool calls.
+- **Lives in a terminal.** Plans the command sequence, checks each result before the next step,
+  and recovers from failures instead of repeating them.
+- **A million tokens.** Paste the repository, not the fragment.
+- **Genuinely multimodal.** Images and video are first-class inputs — read a diagram, a UI
+  screenshot, a stack-trace screenshot, or a whiteboard photo mid-task.
+- **Fast decode for its size.** Hybrid linear+full attention (full every 4th layer) plus an MTP
+  head for self-speculative decoding.
+- **Direct.** Reduced refusal behaviour: it answers the question you asked. See
+  [responsible use](#limitations--responsible-use).
+- **Open weights.** Apache-2.0, `transformers`-native.
+
+## Model overview
+
+| | |
+|---|---|
+| **Parameters** | 27.8B dense (all active) |
+| **Modalities** | text, image, video -> text |
+| **Context window** | 1,048,576 tokens (YaRN + Dual Chunk Attention) |
+| **Attention** | hybrid linear + full attention (full every 4th layer) |
+| **Decoding** | MTP head included (self-speculative decoding) |
+| **Precision** | bfloat16 |
+| **Architecture** | Qwen3.8 dense (27B) + native vision encoder |
+| **License** | Apache-2.0 |
+| **Library** | 🤗 `transformers` (`AutoModelForImageTextToText`) |
+
+The family: [Pro (35B-A3B MoE)](https://huggingface.co/vectionlabs/Salience-1.5-Pro) ·
+[Flash (30B-A3B MoE)](https://huggingface.co/vectionlabs/Salience-1.5-Flash) ·
+**27B R5 (dense)** ·
+[Nano (9B dense)](https://huggingface.co/vectionlabs/Salience-1.5-Nano)
+
+## Capabilities
+
+- **Code & SWE execution** — runnable code, repo-scale edits, methodical debugging, robust backends.
+- **Terminal & agentic work** — multi-step planning, tool orchestration, long-horizon task execution.
+- **Deep reasoning** — structured, inspectable chains of thought for hard, multi-step problems.
+- **Multimodal perception** — diagrams, screenshots, documents, and video as first-class inputs.
+
+## Tool calling
+
+The model emits **XML-style tool calls** (`<tool_call><function=...><parameter=...>`), parsed
+natively by vLLM / SGLang tool parsers for this model family, and by `llama-server --jinja`.
+Provide tool schemas via the chat template `tools` argument.
+
+## Intended use
+
+Salience 27B R5 targets **software engineering, coding agents, and technical research**:
+
+- Code generation, explanation, debugging, review, and repo-scale tasks.
+- Terminal / tool-using agent workflows (CLI agents, browsing, ML engineering, DevOps).
+- Backend and systems design, infrastructure-as-code.
+- Step-by-step reasoning and quantitative problem solving.
+- Screenshot / diagram / document understanding inside engineering workflows.
+
+It is **not** intended for high-stakes decisions without human review, nor as a source of truth
+for medical, legal, or financial advice.
+
+## What ships here
+
+18 safetensors shards, 8 JSON configs, a chat template, this README and a banner.
+**No pickle files, no `.bin`, no `.py`.** Safetensors is a flat tensor container with no
+mechanism for executing code, which is why it replaced pickle — the file list is on this
+page and you do not have to take my word for any of it.
+
+Weights cannot run anything on their own; a harness runs things. If you hand any model
+shell access inside an agent loop that is your risk surface, and it is identical here and
+for the stock base model.
+
+## Quickstart
+
+```python
+from transformers import AutoModelForImageTextToText, AutoProcessor
+import torch
+
+repo = "vectionlabs/Salience-27B-R5"
+proc = AutoProcessor.from_pretrained(repo)
+model = AutoModelForImageTextToText.from_pretrained(
+    repo, dtype="auto", device_map="auto"
+)
+
+messages = [{
+    "role": "user",
+    "content": [{"type": "text", "text": "Implement an LRU cache in Python with O(1) get/put."}],
+}]
+text = proc.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = proc(text=[text], return_tensors="pt").to(model.device)
+out = model.generate(**inputs, max_new_tokens=2048)
+print(proc.batch_decode(out[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)[0])
+```
+
+Requires a recent `transformers` (>= 5.8). Vision works the same way with
+`{"type": "image", "image": ...}` content items.
+
+## Running it locally
+
+| build | format | notes |
+|---|---|---|
+| [bartowski/vectionlabs_Salience-27B-R5-GGUF](https://huggingface.co/bartowski/vectionlabs_Salience-27B-R5-GGUF) | GGUF | full range of sizes |
+| [mradermacher/Salience-27B-R5-i1-GGUF](https://huggingface.co/mradermacher/Salience-27B-R5-i1-GGUF) | GGUF | imatrix |
+| [mradermacher/Salience-27B-R5-GGUF](https://huggingface.co/mradermacher/Salience-27B-R5-GGUF) | GGUF | static |
+| [McG-221/Salience-27B-R5-mlx-8Bit](https://huggingface.co/McG-221/Salience-27B-R5-mlx-8Bit) | MLX | Apple silicon |
+
+Thanks to everyone who built those.
+
+```bash
+llama-server -m Salience-27B-R5-Q4_K_M.gguf \
+  --jinja --reasoning-format deepseek \
+  -c 32768 -ngl 999
+```
+
+`--jinja` is not optional for agent use: it applies the model's own chat template, which is
+what turns XML tool calls into proper OpenAI-style `tool_calls` — and what makes the reasoning
+defaults above take effect. Without it you get malformed calls and stock behaviour.
+
+This is a **dense** model, so standard quant intuition applies: **Q4_K_M and up** hold quality
+well; use Q5_K_M/Q6_K when VRAM allows. (The MoE tiers of the family need Q5/Q6 minimum — that
+constraint does *not* apply here.) Keep the MTP layers if your quant includes them: they enable
+self-speculative decoding for free extra speed.
+
+## Long context
+
+Ships with YaRN (`factor 4.0`, `original_max_position_embeddings 262144`) and a
+`dual_chunk_attention_config` block. Static YaRN taxes short prompts slightly; that is the cost
+of having the full window available by default. vLLM and SGLang read the DCA block,
+`transformers` ignores it.
+
+## Prompting tips
+
+- **Let it think.** No "think step by step" — reasoning is native. Reach for `reasoning_effort`
+  instead of prompt scaffolding.
+- **Give it the repo.** A million tokens: paste whole files or repositories, not fragments.
+- **Agentic loops.** Use `--jinja` with llama-server (or vLLM/SGLang parsers) so XML tool calls
+  become proper OpenAI-style `tool_calls`.
+- **Vision mid-task.** Screenshots of stack traces and UI states work as debugging inputs.
+
+## Benchmarks
+
+**None have been run yet.** Not withheld — not run.
+
+When they exist they will be against a plain Qwen3.8-27B baseline on the same harness and
+the same day, because a number without its baseline is not a measurement. Until then, treat
+everything above as a description of what this model was built to do rather than proof that
+it does it.
+
+## Limitations & responsible use
+
+- May hallucinate APIs or facts under ambiguity; verify critical output.
+- Review generated code before running it, especially anything touching production systems.
+- **Reduced refusal behaviour.** There is no content filter in the weights and no system-level
+  guardrail — the model will attempt requests a stock model declines, and it will not decline
+  on your behalf. Whatever policy your deployment needs is yours to add at the application
+  layer. You are responsible for what you generate and for complying with the law where you
+  operate.
+- `medium` reasoning by default means shorter chains on genuinely hard problems than a model
+  pinned to maximum effort. Pass `reasoning_effort="xhigh"` when the problem deserves it.
+
+---
+
+<sub>Built on Qwen3.8 (Apache-2.0).</sub>
+<sub>Build with love by the vectionlabs' team (Apache-2.0).</sub>
+
+<div align="center"><sub>© 2026 Vection Labs</sub></div>