Add files using upload-large-folder tool

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/scratch/llama405b/base/base",
+  "_name_or_path": "meta-llama/Meta-Llama-3.1-405B",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -7,6 +7,7 @@
   "attention_dropout": 0.0,
   "bos_token_id": 128000,
   "eos_token_id": 128001,
+  "head_dim": 128,
   "hidden_act": "silu",
   "hidden_size": 16384,
   "initializer_range": 0.02,
@@ -16,16 +17,16 @@
   "model_type": "llama",
   "num_attention_heads": 128,
   "num_hidden_layers": 126,
-  "num_key_value_heads": 16,
+  "num_key_value_heads": 8,
   "pretraining_tp": 1,
   "quantization_config": {
     "_load_in_4bit": true,
     "_load_in_8bit": false,
     "bnb_4bit_compute_dtype": "bfloat16",
-    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_storage": "bfloat16",
     "bnb_4bit_quant_type": "nf4",
     "bnb_4bit_use_double_quant": true,
-    "llm_int8_enable_fp32_cpu_offload": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
     "llm_int8_has_fp16_weight": false,
     "llm_int8_skip_modules": null,
     "llm_int8_threshold": 6.0,
@@ -36,15 +37,15 @@
   "rms_norm_eps": 1e-05,
   "rope_scaling": {
     "factor": 8.0,
-    "low_freq_factor": 1.0,
     "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
     "original_max_position_embeddings": 8192,
     "rope_type": "llama3"
   },
   "rope_theta": 500000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.43.0.dev0",
+  "transformers_version": "4.45.0.dev0",
   "use_cache": true,
   "vocab_size": 128256
 }

generation_config.json

@@ -1,7 +1,6 @@
 {
   "_from_model_config": true,
   "bos_token_id": 128000,
-  "do_sample": true,
   "eos_token_id": 128001,
-  "transformers_version": "4.42.4"
+  "transformers_version": "4.45.0.dev0"
 }

model-00002-of-00044.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a87128d152806d9d54a15ac1b5b218be1f0ec70ac650ac896a980dbb9a9b54d2
-size 4985431347
+oid sha256:a565658eef5598c6141e2ac1897141ff92dee9b43ea651ee58172e136ba87e66
+size 4933502323

model-00005-of-00044.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd44523edfac729bf6c321e69081cad8412d27817327e962d5f67dc0dcebc67b
-size 4985431449
+oid sha256:4859e22c88e1e34fb1bd28c1c2a8d421a4ad05b5df5d47b03cc90c82b2043a2d
+size 4933502417

model-00006-of-00044.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:157fbf41f01981bb4f81d42d156bcb5339336a393f26da86fc30a8a5ce3c881a
-size 4985431485
+oid sha256:67e5e02b03b26baf1133d14373ad98680459746ca34e715cbf8847eae66c352f
+size 4933502461

model-00013-of-00044.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:da1c080ebcae50346496ea97773124313a75534124f71cf4a3ec0a56bde7cd43
-size 4985431481
+oid sha256:a7e96999cebcd74c17c133be8e5d20e7e63b9522c36536b178ab7bcd26c7032a
+size 4933502458

model.safetensors.index.json

@@ -1,6 +1,6 @@
 {
   "metadata": {
-    "total_size": 217792871633
+    "total_size": 215611849938
   },
   "weight_map": {
     "lm_head.weight": "model-00044-of-00044.safetensors",

tokenizer.json

@@ -2329,10 +2329,69 @@
     ]
   },
   "post_processor": {
-    "type": "ByteLevel",
-    "add_prefix_space": true,
-    "trim_offsets": false,
-    "use_regex": true
+    "type": "Sequence",
+    "processors": [
+      {
+        "type": "ByteLevel",
+        "add_prefix_space": true,
+        "trim_offsets": false,
+        "use_regex": true
+      },
+      {
+        "type": "TemplateProcessing",
+        "single": [
+          {
+            "SpecialToken": {
+              "id": "<|begin_of_text|>",
+              "type_id": 0
+            }
+          },
+          {
+            "Sequence": {
+              "id": "A",
+              "type_id": 0
+            }
+          }
+        ],
+        "pair": [
+          {
+            "SpecialToken": {
+              "id": "<|begin_of_text|>",
+              "type_id": 0
+            }
+          },
+          {
+            "Sequence": {
+              "id": "A",
+              "type_id": 0
+            }
+          },
+          {
+            "SpecialToken": {
+              "id": "<|begin_of_text|>",
+              "type_id": 1
+            }
+          },
+          {
+            "Sequence": {
+              "id": "B",
+              "type_id": 1
+            }
+          }
+        ],
+        "special_tokens": {
+          "<|begin_of_text|>": {
+            "id": "<|begin_of_text|>",
+            "ids": [
+              128000
+            ],
+            "tokens": [
+              "<|begin_of_text|>"
+            ]
+          }
+        }
+      }
+    ]
   },
   "decoder": {
     "type": "ByteLevel",