clone from WindyWord/translate-pt-en via HF->HF (ADR-039 Phase C gap-fill)

.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 +1,53 @@
+---
+license: cc-by-4.0
+tags:
+- translation
+- marian
+- windyword
+- portuguese
+- english
+language:
+- pt
+- en
+library_name: transformers
+pipeline_tag: translation
+---
+
+# WindyWord.ai Translation — Portuguese → English
+
+**Translates Portuguese (`pt`) text into English (`en`).**
+
+## Quality & Model Lineage
+
+- **Tier:** Standard ⭐⭐⭐⭐
+- **Base model:** [Helsinki-NLP/opus-mt-ROMANCE-en](https://huggingface.co/Helsinki-NLP/opus-mt-ROMANCE-en) — 68k+ downloads, multilingual Romance-source model
+- **Architecture:** MarianMT
+- **Note:** Helsinki-NLP does not publish a dedicated Portuguese-only `opus-mt-pt-en` model. Their canonical path for Portuguese-source translation is the multilingual `opus-mt-ROMANCE-en`, which handles Portuguese alongside Italian, Spanish, French, Catalan, and Romanian as source languages. This WindyWord repo wraps the same underlying weights with a Portuguese-specific repo name to satisfy installer / consumer code that expects a `translate-pt-en` artifact.
+
+For sibling Romance-source pairs use [WindyWord/translate-ROMANCE-en](https://huggingface.co/WindyWord/translate-ROMANCE-en) directly (same underlying model).
+
+## Available Variants
+
+| Variant | Description |
+|---|---|
+| `lora/` | **WindyStandard** — production baseline. Helsinki-NLP/opus-mt-ROMANCE-en weights wrapped under the WindyWord pt-en namespace. |
+
+## Quick Usage
+
+```python
+from transformers import MarianMTModel, MarianTokenizer
+tokenizer = MarianTokenizer.from_pretrained("WindyWord/translate-pt-en", subfolder="lora")
+model = MarianMTModel.from_pretrained("WindyWord/translate-pt-en", subfolder="lora")
+```
+
+## Commercial Use
+
+Visit [windyword.ai](https://windyword.ai) for apps and API access.
+
+---
+
+## Provenance & License
+
+Weights derived from [Helsinki-NLP/opus-mt-ROMANCE-en](https://huggingface.co/Helsinki-NLP/opus-mt-ROMANCE-en) under CC-BY-4.0. Wrapped as a Portuguese-specific entry by WindyWord 2026-04-28 to support installer expectations for a `translate-pt-en` repo. Underlying model is identical to `WindyWord/translate-ROMANCE-en`.
+
+*Certified by Opus 4.6 Opus-Claw (Dr. C) on Veron-1 (RTX 5090, Mt Pleasant SC).*

lora/config.json

@@ -0,0 +1,59 @@
+{
+  "_name_or_path": "/tmp/Helsinki-NLP/opus-mt-ROMANCE-en",
+  "activation_dropout": 0.0,
+  "activation_function": "swish",
+  "add_bias_logits": false,
+  "add_final_layer_norm": false,
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bad_words_ids": [
+    [
+      65000
+    ]
+  ],
+  "bos_token_id": 0,
+  "classif_dropout": 0.0,
+  "classifier_dropout": 0.0,
+  "d_model": 512,
+  "decoder_attention_heads": 8,
+  "decoder_ffn_dim": 2048,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 65000,
+  "decoder_vocab_size": 65001,
+  "dropout": 0.1,
+  "encoder_attention_heads": 8,
+  "encoder_ffn_dim": 2048,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 0,
+  "forced_eos_token_id": 0,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2"
+  },
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2
+  },
+  "max_length": 512,
+  "max_position_embeddings": 512,
+  "model_type": "marian",
+  "normalize_before": false,
+  "normalize_embedding": false,
+  "num_beams": 4,
+  "num_hidden_layers": 6,
+  "pad_token_id": 65000,
+  "scale_embedding": true,
+  "share_encoder_decoder_embeddings": true,
+  "static_position_embeddings": true,
+  "transformers_version": "4.22.0.dev0",
+  "use_cache": true,
+  "vocab_size": 65001
+}

lora/generation_config.json

@@ -0,0 +1,16 @@
+{
+  "bad_words_ids": [
+    [
+      65000
+    ]
+  ],
+  "bos_token_id": 0,
+  "decoder_start_token_id": 65000,
+  "eos_token_id": 0,
+  "forced_eos_token_id": 0,
+  "max_length": 512,
+  "num_beams": 4,
+  "pad_token_id": 65000,
+  "renormalize_logits": true,
+  "transformers_version": "4.32.0.dev0"
+}

lora/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d77bbbd43a214959e027ffc8713fbe31f8609d14827fba645f1361ca20a6f3a
+size 312087009

lora/source.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:baeab21ec00d0b490382b82499ef0348235c5a8e75de28aec8290adf62b007c4
+size 800087

lora/target.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:121d4e0a20cda2052945d790d335f421f7bea7fa6b0ff7fbc9112eed5f12036f
+size 779494

lora/tokenizer_config.json

@@ -0,0 +1 @@
+{"target_lang": "en", "source_lang": "fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la"}