Add Parable banner and match the series card format

README.md

@@ -1,59 +1,80 @@
 ---
-license: apache-2.0
 base_model: AnkitAI/Parable-Qwen3-4B-Claude-Fable-5
+base_model_relation: quantized
+datasets:
+- Glint-Research/Fable-5-traces
+- Roman1111111/gpt5.5-terminal
+license: apache-2.0
+language:
+- en
+pipeline_tag: text-generation
+library_name: mlx
 tags:
 - mlx
 - apple-silicon
-- qwen3
+- 4bit
+- quantized
+- qlora
 - agentic
 - coding
-- lora
-library_name: mlx
-pipeline_tag: text-generation
+- reasoning
+- thinking
+- claude
+- qwen3
 ---
 
-# Parable-Qwen3-4B-Claude-Fable-5 — MLX 4-bit
+# Parable-Qwen3-4B-Claude-Fable-5-MLX-4bit
+
+<picture>
+  <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/ankit-aglawe/parable-assets/main/parable_header_dark.png">
+  <img alt="Parable" src="https://raw.githubusercontent.com/ankit-aglawe/parable-assets/main/parable_header.png">
+</picture>
 
-Apple Silicon build of [Parable-Qwen3-4B-Claude-Fable-5](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5),
-a Qwen3-4B fine-tuned on execution-verified agent traces.
+**Apple Silicon build of Parable-Qwen3-4B: 2.1 GB at 4.501 bits per weight, running natively on MLX with no llama.cpp in the way.**
 
-**2.1 GB, 4.501 bits per weight.** Runs on any M-series Mac with room to spare.
+A 4-bit MLX quantisation of [AnkitAI/Parable-Qwen3-4B-Claude-Fable-5](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5), a Qwen3-4B fine-tune trained on real multi-step agent sessions: planning, tool use, and `<think>` reasoning captured from actual Claude Fable 5 and GPT-5.5 agent work, not synthetic Q&A. Fits comfortably on any M-series Mac.
 
-## Use it
+## Usage
 
 ```bash
 pip install mlx-lm
+```
+
+```bash
 mlx_lm.generate --model AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-MLX-4bit \
-  --prompt "Write a Python function that retries an HTTP call with backoff."
+  --prompt "Write a Python function that retries an HTTP request with exponential backoff."
 ```
 
-Or in Python:
+Or from Python:
 
 ```python
 from mlx_lm import load, generate
+
 model, tokenizer = load("AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-MLX-4bit")
-print(generate(model, tokenizer, prompt="...", max_tokens=512))
+messages = [{"role": "user", "content": "Write a Python function that retries an HTTP request with exponential backoff."}]
+prompt = tokenizer.apply_chat_template(messages, add_generation_prompt=True)
+print(generate(model, tokenizer, prompt=prompt, max_tokens=512))
 ```
 
-## What it is
+## Recipe
 
-Same weights as the source model, quantised to 4-bit for MLX. The recipe behind
-it is v3.1: LoRA on agent traces plus a replay mix, completion-only loss, two
-seeds souped, then merged into the base at scale 0.6 to limit drift.
+v3.1: LoRA on agent traces with a replay mix to limit forgetting, completion-only loss so the model trains on answers rather than prompts, two seeds souped, then merged into the base at scale 0.6 to bound drift from the original weights.
 
-Measured on the 4B, base against tuned in one session on one harness:
+Measured on the full-precision 4B, base against tuned, in one session on one harness:
 
 | | base | v3.1 |
 |---|---|---|
 | HumanEval+ | 0.616 | **0.683** |
 | MBPP+ | 0.603 | **0.638** |
 
-Those numbers are from the full-precision model. Quantisation to 4 bits costs
-some accuracy; they are the ceiling, not a promise for this build.
+Those are the full-precision numbers. Quantising to 4 bits costs accuracy that this table does not measure, so treat them as the ceiling for this build rather than a claim about it.
 
 ## Other formats
 
-- [GGUF](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF) — llama.cpp, LM Studio, Ollama
-- [safetensors](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5) — transformers
+| format | repo | for |
+|---|---|---|
+| GGUF | [Parable-Qwen3-4B-Claude-Fable-5-GGUF](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-GGUF) | llama.cpp, LM Studio, Ollama |
+| MLX 8-bit | [Parable-Qwen3-4B-Claude-Fable-5-MLX-8bit](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5-MLX-8bit) | Apple Silicon, closer to source |
+| safetensors | [Parable-Qwen3-4B-Claude-Fable-5](https://huggingface.co/AnkitAI/Parable-Qwen3-4B-Claude-Fable-5) | transformers |
 
-Apache-2.0, same as the base.
+Apache-2.0, inherited from the base model.