Upload folder using huggingface_hub

.gitattributes

@@ -39,3 +39,6 @@ lora/source.spm filter=lfs diff=lfs merge=lfs -text
 lora/target.spm filter=lfs diff=lfs merge=lfs -text
 herm0/source.spm filter=lfs diff=lfs merge=lfs -text
 herm0/target.spm filter=lfs diff=lfs merge=lfs -text
+ct2/pytorch_model_int8.bin.broken filter=lfs diff=lfs merge=lfs -text
+ct2/source.spm filter=lfs diff=lfs merge=lfs -text
+ct2/target.spm filter=lfs diff=lfs merge=lfs -text

PATIENT_FILE.md

@@ -0,0 +1,26 @@
+# Patient File: windy-pair-is-eo
+**Created:** 2026-03-26T16:17:34.575912+00:00
+**Fleet:** Windy Pro Translation Models
+
+
+## INT8 Quantization — 2026-03-26T16:17:34.575912+00:00
+**Doctor:** Dr. B (Herm Zero)
+**Procedure:** CTranslate2 INT8 quantization of GOATIA/Herm0 improved model
+**Source:** herm0/ (improved model)
+**Output:** herm0-ct2/ (INT8 quantized)
+**Status:** FAILED
+**Error:** MarianMTModel.__init__() got an unexpected keyword argument 'dtype'
+**Recommendation:** Investigate error, may need manual re-export
+
+
+## INT8 Quantization — 2026-03-26T16:20:47.270128+00:00
+**Doctor:** Dr. B (Herm Zero)
+**Procedure:** CTranslate2 INT8 quantization of GOATIA/Herm0 improved model
+**Source:** herm0/ (improved model)
+**Output:** herm0-ct2/ (INT8 quantized)
+**Status:** SUCCESS
+**Original size:** 183.8 MB
+**INT8 size:** 47.5 MB
+**Compression:** 3.87x
+**Load test:** passed
+**Recommendation:** INT8 variant ready for production deployment

ct2/config.json

@@ -0,0 +1,55 @@
+{
+  "activation_dropout": 0.0,
+  "activation_function": "swish",
+  "add_bias_logits": false,
+  "add_final_layer_norm": false,
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "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": 7501,
+  "decoder_vocab_size": 7502,
+  "dropout": 0.1,
+  "dtype": "float32",
+  "encoder_attention_heads": 8,
+  "encoder_ffn_dim": 2048,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 0,
+  "extra_pos_embeddings": 7502,
+  "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": null,
+  "max_position_embeddings": 512,
+  "model_type": "marian",
+  "normalize_before": false,
+  "normalize_embedding": false,
+  "num_beams": null,
+  "num_hidden_layers": 6,
+  "pad_token_id": 7501,
+  "scale_embedding": true,
+  "share_encoder_decoder_embeddings": true,
+  "static_position_embeddings": true,
+  "transformers_version": "4.57.3",
+  "use_cache": true,
+  "vocab_size": 7502
+}

ct2/generation_config.json

@@ -0,0 +1,16 @@
+{
+  "bad_words_ids": [
+    [
+      7501
+    ]
+  ],
+  "bos_token_id": 0,
+  "decoder_start_token_id": 7501,
+  "eos_token_id": 0,
+  "forced_eos_token_id": 0,
+  "max_length": 512,
+  "num_beams": 6,
+  "pad_token_id": 7501,
+  "renormalize_logits": true,
+  "transformers_version": "4.57.3"
+}

ct2/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:980184e435469d60f8a2623707bdbfdefe73c022948b574ba6ca2b7f66c4e9ab
+size 191977136

ct2/pytorch_model_int8.bin.broken

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28bc119079d7b09a686357607171750cb9cc25e81c84fa3fcf5ce0d713577d0b
+size 65922570

ct2/source.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51a5ad1e836e06c34ab3c4e818d5e4e7cb7159ae63a2932dbd9007cac4e601d1
+size 303988

ct2/special_tokens_map.json

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

ct2/target.spm

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9ab02ecc63d921c41ae6382a55c6670fa88c5dd1dfb035ee754347dc318fb89
+size 304111

ct2/tokenizer_config.json

@@ -0,0 +1,38 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7501": {
+      "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": "isl",
+  "sp_model_kwargs": {},
+  "target_lang": "epo",
+  "tokenizer_class": "MarianTokenizer",
+  "unk_token": "<unk>"
+}

herm0-ct2/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>"
+}

herm0-ct2/model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3615d9705453750600a77459b1a388de3669a93c8bd8e3173bbe00b6c6087fba
+size 49668155