Add files using upload-large-folder tool

chat_template.jinja

@@ -100,4 +100,4 @@
 
 {%- if add_generation_prompt -%}
     {{ '<|Assistant|>' }}{{ '<|end_of_thought|>' }}
-{%- endif -%}
+{%- endif -%}

config.json

@@ -10,9 +10,11 @@
         "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
     },
     "bos_token_id": 0,
+    "dtype": "bfloat16",
     "eos_token_id": 1,
     "ep_size": 1,
     "first_k_dense_replace": 1,
+    "head_dim": 64,
     "hidden_act": "silu",
     "hidden_size": 2048,
     "initializer_range": 0.02,
@@ -31,21 +33,11 @@
     "num_hidden_layers": 40,
     "num_key_value_heads": 32,
     "num_nextn_predict_layers": 1,
+    "pretraining_tp": 1,
     "q_lora_rank": 1536,
+    "qk_head_dim": 192,
     "qk_nope_head_dim": 128,
     "qk_rope_head_dim": 64,
-    "rms_norm_eps": 1e-06,
-    "rope_theta": 32000000,
-    "routed_scaling_factor": 2.5,
-    "scoring_func": "sigmoid",
-    "tie_word_embeddings": false,
-    "topk_group": 1,
-    "topk_method": "noaux_tc",
-    "torch_dtype": "bfloat16",
-    "transformers_version": "4.44.2",
-    "use_cache": true,
-    "v_head_dim": 128,
-    "vocab_size": 129280,
     "quantization_config": {
         "config_groups": {
             "group_0": {
@@ -58,7 +50,7 @@
                     "actorder": null,
                     "block_structure": null,
                     "dynamic": false,
-                    "group_size": 32,
+                    "group_size": 64,
                     "num_bits": 4,
                     "observer": "minmax",
                     "observer_kwargs": {},
@@ -78,5 +70,18 @@
         "kv_cache_scheme": null,
         "quant_method": "compressed-tensors",
         "quantization_status": "compressed"
-    }
+    },
+    "rms_norm_eps": 1e-06,
+    "rope_interleave": true,
+    "rope_scaling": null,
+    "rope_theta": 32000000,
+    "routed_scaling_factor": 2.5,
+    "scoring_func": "sigmoid",
+    "tie_word_embeddings": false,
+    "topk_group": 1,
+    "topk_method": "noaux_tc",
+    "transformers_version": "4.57.3",
+    "use_cache": true,
+    "v_head_dim": 128,
+    "vocab_size": 129280
 }

model-1-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:406da2e446429c6926744227f90f55b1d57a36c37163c63af496422ec04677ca
+oid sha256:d4ff3a9595c500b44c308dcb352de70289268867c1cb95398c991b6bda655288
 size 140785016

model-10-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a9a14526f144c3c70dea5f4129e5e2db227b232efce2392c525142223fd4788
-size 818455784
+oid sha256:726a23a975f03162a5fbefa721fb5f06489aaf7a3a6932b7e0ef91e3a1aa7a89
+size 742957112

model-11-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:591bea12f9179342ac8473df01ded18fa791e0434864a91a70606511adfc4d9c
-size 818458104
+oid sha256:bd640c160281bd25ee639d7827d8ef1973b88ebdb0bd6e2b4707c57111235329
+size 742959432

model-12-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca31c4dd871d00391cd4c1c5f0e00579d9adc4a2eced536b0db1a1f2d94c1307
-size 818458104
+oid sha256:a52a909c7291bacc83be6583a74fa636bc1f80428bd99a4cdd5d9e04b1fab9db
+size 742959432

model-13-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:96dc2b62833028c2ceb4a9836afada274188ea6ab6af1d100bf809c86f270567
-size 818458104
+oid sha256:acc59d9a7e9c461521d0cfe97a17be25f9964139437c877d2ee9b06eadfe0635
+size 742959432

model-14-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:308faf2cb085ca476e58a64f4c556170c41ae9c43e498739ec8a399bd6abd9e1
-size 818458104
+oid sha256:2ba42417e9ed554dc63689c800b5ac05e0022cd82577e9eccb39336733915e96
+size 742959432

model-15-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e938314f5cadf447946ded702943747039b22fa6c1e8a496a78ad11f4f8f9f53
-size 818458104
+oid sha256:547cf7c15e82f400a52d008133236b614f619db00c7ab747ba6457294a36da24
+size 742959432

model-16-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:02ce3eee43cd09f0075e517c6cca57e6833ca330a5eeb751775b3297150485c3
-size 818458104
+oid sha256:e0d7632f46d804d3f8fed17525ad22a79b4304366c0e27ba0c943dad384c9e45
+size 742959432

model-17-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4e981b2b5a10bafbf87279c6ae2f1bb5f2786fd239cff1fe6057a3221a128645
-size 818458104
+oid sha256:5f18a1c3ef7485c1d949a4766d8dcec32ac043df1c8bf911758718156fccbc5b
+size 742959432

model-18-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d463ff3ee96fde06710799c08054b06991301c2331c998fb0c4d2dd964783698
-size 818458104
+oid sha256:cfbe35220cb9898a22a4fdf2be7402f031e1dbac742a65b317d26509c5764b34
+size 742959432

model-19-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:894bd182b39e4d42b80d9cc40ac5a38b803f1754aa270867424cc35abbb8e509
-size 818458104
+oid sha256:0a10861e1ffc82ac0f4461fd581b6b401ee5eecbc655c5a8c82b985f5c5fdd97
+size 742959432

model-2-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6bdd822798f1569b0e201be7819ef9097895546e57883b03d0ac9e307976818d
-size 818455784
+oid sha256:350f55333a55f367cc9c8dac8d08def69eeeb2b87fb7e883095a242a16ca6485
+size 742957112

model-20-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fb3862ffbaf77d91c45195fd3febfe07c7651c624db323c5df09637bf419cb6
-size 818458104
+oid sha256:59c28ace24b7d7f918eb6d0b1aee178da684e9e0e87007f90682ca6c227fa7cd
+size 742959432

model-21-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a755db4f21c8ef10971a548d8f9d699fecef3fef668fe94e9326761d8e10bf04
-size 818458104
+oid sha256:f4b94080b3a18841a127d565dae1be4e555f3a6108ff4c8cbcd67cbe550739d1
+size 742959432

model-22-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:de93a5f68c185218e91ab1286409dbc97465de8e8fc80dac9faa2dd6bc57de81
-size 818458104
+oid sha256:19ddcdcafae2e4f809db86a353e600944e33e9979be9f3c7b0a8d2f010f427f1
+size 742959432

model-23-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:751a31400e286a22e9d46c1a20f03adbdc4476f936ba8507b2af033aa66fc38f
-size 818458104
+oid sha256:41f14a2707f5d5f2f9c9c66dcae3e3804ead083fa941776e3f897d7e16252987
+size 742959432

model-24-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6ca8b8535f50e822f8cd57a8b87e9f5a6f231de0f562a2717ea744cc5f617835
-size 818458104
+oid sha256:53938e077028c98ec68950310513de8fb6fb7123d6d92b6cc7581f728eef323e
+size 742959432

model-25-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1ddedcf14d393e3ec36c223f9bd205a440ad6b5105fa140b23532f5428ade9e7
-size 818458104
+oid sha256:f5410eb0f47bddba92dc7f167d649aaf1110674679ca48e07fe64e7eeb6f6197
+size 742959432

model-26-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:830d845fc512e2a790e6798fc203871c056ea612eddddb559ccd36752d93459b
-size 818458104
+oid sha256:21bc943a8428b23d0054fd1e10b7506d5d84ec137b3aedae7a7a4589522c9b51
+size 742959432

model-27-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3298df21b79112125e7ad095172e9c562d5581f59894b73b7d43f1745f22d89
-size 818458104
+oid sha256:554e73d388f048d7619a13d0f79c1a80ca6e9279a8073ff8dae186df526e311f
+size 742959432

model-28-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3811c01f6d50477a07e586492508019ecc7883f1a69ba7e7a8e9661f3b4395b2
-size 818458104
+oid sha256:3fc9b275e1d16621ba95656c4c5dd1a47ab822aef33ec78d398613851bb43efc
+size 742959432

model-29-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b85ffaf5973be8fa030171616de0e61d9965766c923762d7143125033940bcb7
-size 818458104
+oid sha256:7c8c41d732ca08ee9a2efa0ff166eb7e8b3537e371be3974e26e3ea1fc7d088a
+size 742959432

model-3-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0cd9ca1bebbda88f726d5cc5db88c89a707f4a9ff4a1bb686f77ec3d51bbb66f
-size 818455784
+oid sha256:10b9f33a96d3967e5036ee3904585c6c598f636c30868782a088d8357c647993
+size 742957112

model-30-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0afb7264ac96c3d4f3ecef5bbb963092c43eee3cce8e98cdda06a5ac23edff42
-size 818458104
+oid sha256:9fb82b7519ecf2dd6899f3aa6a9c1b270c0b45c11c2739766fcebf4a520f908a
+size 742959432

model-31-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7adb6a2b2ba85188b0578cff4c99768c9677ae9629590d7a820065e9119dcc0d
-size 818458104
+oid sha256:a783a6038833bda2b6f781d414daebc6e87b341079ebf402d70e408cbbb7e152
+size 742959432

model-32-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:14af1ce73f032d002f99fc22a79d7b9222a619af0f3a0a94341d6812885d9322
-size 818458104
+oid sha256:772c68a4ce2a49aba2a37fef2dc832f93f63456f85eafc260d03ae2cf9b5ed08
+size 742959432

model-33-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f298b903f77900ceaf9beefdae99d2b26bcdfdee7c450a971e2dc5fa9724e73
-size 818458104
+oid sha256:5ed2633aaa725ad82583ebecb9f24c00c67a80a5669c309a9c460cbddc4948c2
+size 742959432

model-34-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:706fdfbe986f6297b8585c053bdecf06357d163f1a967799aa28792a898ad38a
-size 818458104
+oid sha256:7b2eea6d07916ef281c8f0dae32d8d26c7e4d093f6a77a065daf5f97a668de4a
+size 742959432

model-35-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b5b4dafab74c492e188001486016d14c57b9787b948eb3cf6d1a2d1bcc558e20
-size 818458104
+oid sha256:7dadaa7799c2217fe60ffcf866a0bdb8cbf80514fa8651cd55d6b27bd6e3533e
+size 742959432

model-36-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27c762a92d19a55b0deeb065292e86b0800d9bd2b44533077efee5dbb06ce0db
-size 818458104
+oid sha256:b66da6210a4112ebd30d1d27bac0e03534b14a3452fc3fe2f21ad00fc1c121e0
+size 742959432

model-37-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee716e34f3f48b4cd9f9be984ea9adcdd20c5c39c62cde1a96b1e58733a56118
-size 818458104
+oid sha256:c33f20ea2a6cdcc1a6b7ac03019528ba56721b6c521448297744bd4f327731a6
+size 742959432

model-38-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f584b697b83471f0c044ed7bff158ec2d36e3b9464c999ac82dac09aa3e910b
-size 818458104
+oid sha256:b6970a9ac0d9c47ffedd401ce63deff887ac40ee7c2bc67f297ca11384e00d62
+size 742959432

model-39-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5127a9bcfa5b4c8ef37ec0c7b12e418ad6ba566a80c0ce59c331318a6c5f40e2
-size 818458104
+oid sha256:21fa2b399986e7f990426980a475e2633d68db49611718bd23a36190a8da2b08
+size 742959432

model-4-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0801dd78850f5dfe8e861c010adda50c3e26d602d718d5a57ff0245ebb142a8d
-size 818455784
+oid sha256:4298c221ca35e3c9607b0d7ff0411d7abd95a1fa543b961d3696d3850b50ed3e
+size 742957112

model-40-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:67463b6637636bf9d2adb645f56ac1970930ed919832d1f5f987419f5b00b804
-size 818458104
+oid sha256:61742d29c756dd672d4822e0c735649e9e31ab7e25254f99262a76785f30802f
+size 742959432

model-5-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e4342823e80e2e740a97070fec31b98a8b65dd53fd6eff68d22afd421346cd3
-size 818455784
+oid sha256:a5e7a76aa67b9201962fdcb23048229fa07809fc988a278527691ef18125ab1d
+size 742957112

model-6-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37cf69249bd0c3a58e86bfc17674c4b9c512c80ae75737c730c679a5f9a4440e
-size 818455784
+oid sha256:89cb41236f515d0290b0314995ff844244edff51adddb15250d9cd0a0620b376
+size 742957112

model-7-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4b3b758f044f42f2aeb6807d6adbf90b810d72e2a8cb264ab4388a281a4aee9d
-size 818455784
+oid sha256:fd2753a823e2efaf7d072d2bd79df61b71276107e4f6222e7fb0e0f8e27e6cc8
+size 742957112

model-8-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c69250b22fcb82655f84d7cd1fe5095a1144ffe733173aea997f21bda7805857
-size 818455784
+oid sha256:1f6d5145ecd921a7d7d861daf849cb777bf7b8e7567cf5969e353e4e759709f8
+size 742957112

model-9-of-40.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5540d20c8ef9b3af7677341c7fb094833c97b979fb1c354186773b2943f4d3fd
-size 818455784
+oid sha256:9795060b455ea238e1cdb64df5aee4e51d62bae3e51d4c31e579718aedd0d43a
+size 742957112

model-non-layer.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:feab4d0c54e6ce4dfcd72044cd8db156eccb363c23734515ba070ec34001e98f
+oid sha256:9b992b0a3e3ab06a9490d364bf942df3bcf69874dcb9f940e935f674672f09cd
 size 1059066184

model.safetensors.index.json

@@ -1,7 +1,4 @@
 {
-    "metadata": {
-        "total_size": 0
-    },
     "weight_map": {
         "lm_head.weight": "model-non-layer.safetensors",
         "model.embed_tokens.weight": "model-non-layer.safetensors",
@@ -90435,7 +90432,7 @@
         "model.layers.40.self_attn.q_a_layernorm.weight": "mtp-1-of-1.safetensors",
         "model.layers.40.self_attn.q_a_proj.weight": "mtp-1-of-1.safetensors",
         "model.layers.40.self_attn.q_b_proj.weight": "mtp-1-of-1.safetensors",
-        "model.layers.40.shared_head.norm.weight": "mtp-1-of-1.safetensors",
-        "model.layers.40.shared_head.head.weight": "mtp-1-of-1.safetensors"
+        "model.layers.40.shared_head.head.weight": "mtp-1-of-1.safetensors",
+        "model.layers.40.shared_head.norm.weight": "mtp-1-of-1.safetensors"
     }
 }

mtp-1-of-1.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5668c90e7a7cab0ea4573fccc0cd8e4bcdd7aa18e8f9b3b7b8bed169ca0ec1c
+oid sha256:74cb5a9b60987b70d854406b19ed713e781774bada163d3097f808f09c08d7bf
 size 687105960