Upload folder using huggingface_hub

README.md

@@ -1,61 +1,7 @@
 ---
-license: apache-2.0
-base_model: Qwen/Qwen3-0.6B
+language: en
 library_name: mlx
-tags: [mlx, text-simplification, easy-language, accessibility, multilingual, qwen3]
-language: [de, fr, en, es, ar, da, et, fi, hi, it, ja, ko, nl, pt, ru, sk, sv, tr, vi]
+pipeline_tag: text-generation
+tags:
+- mlx
 ---
-
-# Qwen3-0.6B-EasyLanguage (4-bit, MLX)
-
-LoRA fine-tune of **Qwen/Qwen3-0.6B** that rewrites live speech transcripts into easy-to-read
-registers across **20 language locales** — German (Leichte Sprache), French (FALC), English (Easy/Plain English), Spanish (Lectura Fácil), Easy-to-Read Arabic (Inclusion Europe / Information for All), Letlæst (Inclusion Europe ETR), and more — for the
-[Live Linguist](https://github.com/ngoldbla/live-linguist) on-device captioner.
-Each locale follows its own national / European Easy-to-Read or plain-language standard.
-Quantized to 4-bit for Apple-silicon inference via MLX.
-
-It splits run-ons into short sentences, drops disfluencies, keeps names/numbers, and
-stays in the input language. Trained on a lean prompt (no few-shots) so the register is
-internalized — shorter prompts, lower live-caption latency.
-
-## Evaluation (held-out test set; SARI = simplification quality)
-| lang | SARI ft | SARI stock | chrF ft | LID ft | compliance ft |
-| --- | --- | --- | --- | --- | --- |
-| de | 47.89 | 32.98 | 42.26 | 1.0 | 0.99 |
-| fr | 56.93 | 33.31 | 51.19 | 1.0 | 1.0 |
-| es | 59.34 | 36.57 | 54.79 | 1.0 | 0.995 |
-| en | 59.17 | 32.46 | 54.57 | 1.0 | 1.0 |
-| ar | 49.23 | 50.11 | 45.74 | 0.985 | 0.94 |
-| da | 53.12 | 35.58 | 50.24 | 0.995 | 0.94 |
-| et | 49.39 | 30.48 | 52.76 | 1.0 | 1.0 |
-| fi | 50.55 | 35.2 | 56.33 | 1.0 | 0.985 |
-| hi | 52.66 | 35.16 | 45.05 | 0.995 | 0.82 |
-| it | 53.34 | 49.18 | 48.83 | 1.0 | 0.8299 |
-| ja | 8.37 | 8.29 | 40.85 | 1.0 | 0.97 |
-| ko | 49.55 | 43.77 | 40.03 | 1.0 | 0.98 |
-| nl | 51.9 | 38.96 | 48.38 | 0.985 | 0.9 |
-| pt-BR | 55.02 | 37.54 | 52.11 | 1.0 | 0.82 |
-| pt-PT | 57.44 | 38.96 | 58.14 | 1.0 | 0.975 |
-| ru | 50.61 | 41.13 | 49.65 | 1.0 | 0.94 |
-| sk | 50.92 | 40.59 | 46.68 | 1.0 | 0.975 |
-| sv | 53.23 | 40.16 | 55.08 | 0.995 | 0.96 |
-| tr | 51.32 | 31.73 | 49.91 | 0.995 | 0.97 |
-| vi | 59.62 | 41.3 | 58.75 | 1.0 | 0.725 |
-| zh-CN | 8.75 | 8.5 | 51.67 | 0.955 | 0.925 |
-
-## Usage (MLX)
-```python
-from mlx_lm import load, generate
-model, tok = load("ndgold/Qwen3-0.6B-EasyLanguage-4bit")
-msgs = [{"role":"system","content":"<framework system prompt>"},
-        {"role":"user","content":"Original: <utterance>\nRewritten:"}]
-p = tok.apply_chat_template(msgs, add_generation_prompt=True, tokenize=False, enable_thinking=False)
-print(generate(model, tok, prompt=p, max_tokens=96))
-```
-
-## Sources & licenses
-- **Base model:** Qwen3 (Apache-2.0).
-- **German seed data:** [tum-nlp/German4All-Corpus](https://huggingface.co/datasets/tum-nlp/German4All-Corpus) (German Wikipedia, CC BY-SA).
-- **Synthetic pairs (all other languages + de augmentation):** spoken→easy-language pairs generated by Claude (Anthropic) — Leichte Sprache (de), FALC (fr), Easy/Plain English (en), Lectura Fácil (es), and 25 further locales following each language's Easy-to-Read / plain-language standard (Inclusion Europe "Information for All", Selkokieli, Lättläst, やさしい日本語, ISO 24495-1, …). Every pair is filtered by a deterministic validator suite (per-language sentence-length caps, language-ID, fidelity anchoring, number preservation, anti-parroting).
-- Intended for the **Live Linguist** on-device live-caption simplifier. Not a general chatbot.
-

config.json

@@ -5,6 +5,7 @@
     "attention_bias": false,
     "attention_dropout": 0.0,
     "bos_token_id": 151643,
+    "dtype": "bfloat16",
     "eos_token_id": [
         151645,
         151643
@@ -14,6 +15,36 @@
     "hidden_size": 1024,
     "initializer_range": 0.02,
     "intermediate_size": 3072,
+    "layer_types": [
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention",
+        "full_attention"
+    ],
     "max_position_embeddings": 40960,
     "max_window_layers": 28,
     "model_type": "qwen3",
@@ -35,8 +66,7 @@
     "rope_theta": 1000000,
     "sliding_window": null,
     "tie_word_embeddings": true,
-    "torch_dtype": "bfloat16",
-    "transformers_version": "4.51.0",
+    "transformers_version": "4.57.6",
     "use_cache": true,
     "use_sliding_window": false,
     "vocab_size": 151936

generation_config.json

@@ -1,13 +1,13 @@
 {
-    "bos_token_id": 151643,
-    "do_sample": true,
-    "eos_token_id": [
-        151645,
-        151643
-    ],
-    "pad_token_id": 151643,
-    "temperature": 0.6,
-    "top_k": 20,
-    "top_p": 0.95,
-    "transformers_version": "4.51.0"
-}
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "temperature": 0.6,
+  "top_k": 20,
+  "top_p": 0.95,
+  "transformers_version": "4.57.6"
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:314cb1b4bbbfc57edce46292ee4d73c66f95c24068361aaa981f1c222d91d4a3
+oid sha256:f00c3328b59c3650a1a6f5413f5587e26d15c37b6ae5d75bf1214b40b1889c9a
 size 335450548

tokenizer_config.json

@@ -5,21 +5,6 @@
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|im_end|>",
   "errors": "replace",
-  "extra_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>",
-    "<|object_ref_start|>",
-    "<|object_ref_end|>",
-    "<|box_start|>",
-    "<|box_end|>",
-    "<|quad_start|>",
-    "<|quad_end|>",
-    "<|vision_start|>",
-    "<|vision_end|>",
-    "<|vision_pad|>",
-    "<|image_pad|>",
-    "<|video_pad|>"
-  ],
   "is_local": true,
   "local_files_only": false,
   "model_max_length": 131072,