clone from WindyWord/translate-tc-bible-big-mkh-fra_ita_por_spa via local upload (ADR-039 Phase C)

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+lora/source.spm filter=lfs diff=lfs merge=lfs -text
+lora/target.spm filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,84 @@
+---
+license: cc-by-4.0
+tags:
+- translation
+- marian
+- windyword
+- mon-khmer
+- vietnamese
+- khmer
+- mon
+- french-italian-portuguese-spanish
+- french
+- italian
+- portuguese
+- spanish
+language:
+- mkh
+- fra
+- ita
+- por
+- spa
+library_name: transformers
+pipeline_tag: translation
+---
+
+# WindyWord.ai Translation — Mon-Khmer → French/Italian/Portuguese/Spanish
+
+**Translates Mon-Khmer (Vietnamese, Khmer, Mon) → French / Italian / Portuguese / Spanish.**
+
+**Quality Rating: ⭐⭐½  (2.5★ Basic)**
+
+Part of the [WindyWord.ai](https://windyword.ai) translation fleet — 1,800+ proprietary language pairs.
+
+## Quality & Pricing Tier
+
+- **5-star rating:** 2.5★ ⭐⭐½
+- **Tier:** Basic
+- **Composite score:** 52.5 / 100
+- **Rated via:** Grand Rounds v2 — an 8-test stress battery (paragraphs, multi-paragraph, native input, domain stress, edge cases, round-trip fidelity, speed, and consistency checks)
+
+## Available Variants
+
+This repository contains multiple deployment formats. Pick the one that matches your use case:
+
+| Variant | Description |
+|---|---|
+| `lora/` | **WindyStandard** — our proprietary production baseline. Stable, reliable, optimized for GPU inference. |
+| `lora-ct2-int8/` | **WindyStandard · CPU INT8** — CTranslate2 quantized version of WindyStandard. ~25% of the size, 2–4× faster on CPU, no measurable quality loss. |
+
+### Quick usage
+
+**Transformers (PyTorch):**
+```python
+from transformers import MarianMTModel, MarianTokenizer
+tokenizer = MarianTokenizer.from_pretrained("WindyWord/translate-tc-bible-big-mkh-fra_ita_por_spa", subfolder="lora")
+model = MarianMTModel.from_pretrained("WindyWord/translate-tc-bible-big-mkh-fra_ita_por_spa", subfolder="lora")
+```
+
+**CTranslate2 (fast CPU inference):**
+```python
+import ctranslate2
+translator = ctranslate2.Translator("path/to/translate-tc-bible-big-mkh-fra_ita_por_spa/lora-ct2-int8")
+```
+
+## Commercial Use
+
+The WindyWord.ai platform provides:
+- **Mobile apps** (iOS, Android — coming soon)
+- **Real-time voice-to-text-to-translation** pipeline
+- **API access** with premium model quality
+- **Offline deployment** support
+
+Visit [windyword.ai](https://windyword.ai) for apps and commercial API access.
+
+---
+
+## Provenance & License
+
+Weights derived from the OPUS-MT project ([Helsinki-NLP/opus-mt-tc-bible-big-mkh-fra_ita_por_spa](https://huggingface.co/Helsinki-NLP/opus-mt-tc-bible-big-mkh-fra_ita_por_spa)) under CC-BY-4.0. WindyStandard, WindyEnhanced, and WindyScripture variants are proprietary to WindyWord.ai, independently trained and quality-certified via our Grand Rounds v2 test battery.
+
+Licensed CC-BY-4.0 — attribution preserved as required.
+
+*Certified by Opus 4.6 Opus-Claw (Dr. C) on Veron-1 (RTX 5090).*
+*Patient file: [clinic record](https://github.com/sneakyfree/Windy-Clinic/blob/main/translation-pairs/tc-bible-big-mkh-fra_ita_por_spa.json)*

lora-ct2-int8/config.json

@@ -0,0 +1,10 @@
+{
+  "add_source_bos": false,
+  "add_source_eos": false,
+  "bos_token": "<s>",
+  "decoder_start_token": "</s>",
+  "eos_token": "</s>",
+  "layer_norm_epsilon": null,
+  "multi_query_attention": false,
+  "unk_token": "<unk>"
+}

lora-ct2-int8/model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80fa4b12858948589ffc9eef1413122fe7419c0c622eaa4ad55be09531299ec2
+size 241801171

lora/config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "Helsinki-NLP/opus-mt-tc-bible-big-mkh-fra_ita_por_spa",
+  "activation_dropout": 0.0,
+  "activation_function": "relu",
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.0,
+  "d_model": 1024,
+  "decoder_attention_heads": 16,
+  "decoder_ffn_dim": 4096,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 58240,
+  "decoder_vocab_size": 58241,
+  "dropout": 0.1,
+  "encoder_attention_heads": 16,
+  "encoder_ffn_dim": 4096,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 1047,
+  "forced_eos_token_id": null,
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "max_length": null,
+  "max_position_embeddings": 1024,
+  "model_type": "marian",
+  "normalize_embedding": false,
+  "num_beams": null,
+  "num_hidden_layers": 6,
+  "pad_token_id": 58240,
+  "scale_embedding": true,
+  "share_encoder_decoder_embeddings": true,
+  "static_position_embeddings": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.45.1",
+  "use_cache": true,
+  "vocab_size": 58241
+}

lora/generation_config.json

@@ -0,0 +1,16 @@
+{
+  "_from_model_config": true,
+  "bad_words_ids": [
+    [
+      58240
+    ]
+  ],
+  "bos_token_id": 0,
+  "decoder_start_token_id": 58240,
+  "eos_token_id": 1047,
+  "forced_eos_token_id": 1047,
+  "max_length": 512,
+  "num_beams": 4,
+  "pad_token_id": 58240,
+  "transformers_version": "4.45.1"
+}

lora/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd674841409d9050a97ff8162ffa75ed4197d60efdf906087c8cf3b483b000c
+size 944247220

lora/source.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27ee3d1c1328f9681c5fdfcbbc7104b2ebb1c677ece7b77576561ab877906cef
+size 878704

lora/special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "eos_token": "</s>",
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

lora/target.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb3745d5322c9845a7620ea7d3fe009e55cfbddc3b8bbf354be917936bc2ddd2
+size 816024

lora/tokenizer_config.json

@@ -0,0 +1,38 @@
+{
+  "added_tokens_decoder": {
+    "1047": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1305": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "58240": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "model_max_length": 512,
+  "pad_token": "<pad>",
+  "separate_vocabs": false,
+  "source_lang": "mkh",
+  "sp_model_kwargs": {},
+  "target_lang": "fra+ita+por+spa",
+  "tokenizer_class": "MarianTokenizer",
+  "unk_token": "<unk>"
+}