Upload folder using huggingface_hub

.gitattributes

@@ -35,3 +35,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 onnx/model.onnx_data filter=lfs diff=lfs merge=lfs -text
 onnx/model_fp16.onnx_data filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

config.json

@@ -5,7 +5,7 @@
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 151643,
-  "eos_token_id": 151643,
+  "eos_token_id": 151645,
   "hidden_act": "silu",
   "hidden_size": 1536,
   "initializer_range": 0.02,
@@ -17,11 +17,19 @@
   "num_hidden_layers": 28,
   "num_key_value_heads": 2,
   "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
   "rope_theta": 1000000.0,
   "sliding_window": null,
   "tie_word_embeddings": true,
-  "transformers_version": "4.43.4",
+  "torch_dtype": "float32",
+  "transformers_version": "4.46.1",
   "use_cache": true,
   "use_sliding_window": false,
-  "vocab_size": 151936
-}
+  "vocab_size": 151936,
+  "transformers.js_config": {
+    "kv_cache_dtype": {
+      "q4f16": "float16",
+      "fp16": "float16"
+    }
+  }
+}

generation_config.json

@@ -10,5 +10,5 @@
   "temperature": 0.7,
   "top_k": 20,
   "top_p": 0.8,
-  "transformers_version": "4.43.4"
+  "transformers_version": "4.46.1"
 }

onnx/model.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f7da5d0bb5e5b6eba0ba0e9c006fcbc8b670134405ef5e02aaaf738361e2074
-size 1057843
+oid sha256:f8a2891765a8cfb5fbf76235fcb99f84d57a47430f9efeb24ef400e17c91d019
+size 532187

onnx/model.onnx_data

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a2e49301861eea69658adedddda876c537bc8750f5493420bdf3dc66a7a0f93f
-size 6208411648
+oid sha256:0a11022897439e8f65c9882338322897453a05abe3435b5ed38094e128f0c080
+size 7125129580

onnx/model_bnb4.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:11c2db77de0a2f18b499a91f68ddfd3cf238dc50be8039c51675aad208091a05
-size 1705680982
+oid sha256:a22d663bf3915f9c5a111d5be6299e18d22d5f503ce8aaa31231eff11c995779
+size 1819659202

onnx/model_fp16.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:274c6bd20bd1733880f203a64c5d932a912aed948f0a10f3756f3ae9dd37dcd0
-size 1098300
+oid sha256:266034903747b160bbc23f7a8689c69d622a34941b94d54543d5374b1d694009
+size 533991

onnx/model_fp16.onnx_data

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ac8af43dc889f27d014cef1e46a31e3f307e39cf9ac09f1c3716c104573fb16c
-size 3104177152
+oid sha256:69b1fd26da7afc014e57b7a64cef8265bd545809f358798b48e0af784400e6d1
+size 3562564966

onnx/model_int8.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee0ebbf3863b640569243ba3fbf9fb05bcc95ccf488133a329b168bfe347772b
-size 1578954293
+oid sha256:92f3b69db6c85669db5ed18d501a6e47f8aaf849b847a3c27a2014905db1cc17
+size 1795035770

onnx/model_q4.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:63d13a2f96ff0f2ac1b23050aba04f0065ecb29ff44b5d4abad8b3b8004d35ac
-size 1787566590
+oid sha256:3be13b6b30c1c152bd36e5a886a36800d404f56494c74e2fdf1d634d175372e1
+size 1915754790

onnx/model_q4f16.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:54ead74db2218bbea5c79d964494a16c9e53cd76a4e13590dd200542d3322f05
-size 1221878940
+oid sha256:0e33cbfde4be371039daad0032fb3782170442e6901a25d4deabe96117b89882
+size 1344237070

onnx/model_quantized.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee0ebbf3863b640569243ba3fbf9fb05bcc95ccf488133a329b168bfe347772b
-size 1578954293
+oid sha256:92f3b69db6c85669db5ed18d501a6e47f8aaf849b847a3c27a2014905db1cc17
+size 1795035770

onnx/model_uint8.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec159ffadd7b0c820f8ce620dae4512c6062c582acfb4be34e9d293b1fa2367f
-size 1578954390
+oid sha256:967d74ee65ca2ee796697b90b2875f24255ead1ea9d2861cbafc592987861ba6
+size 1795035868