Upload with huggingface_hub

config.json

@@ -1 +1 @@
-{"architectures": ["LLaMAForCausalLM"], "bos_token_id": 0, "eos_token_id": 1, "hidden_act": "silu", "hidden_size": 4096, "intermediate_size": 11008, "initializer_range": 0.02, "max_sequence_length": 2048, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "pad_token_id": -1, "rms_norm_eps": 1e-06, "torch_dtype": "float16", "transformers_version": "4.27.0.dev0", "use_cache": true, "vocab_size": 32000}
+{"architectures": ["LlamaForCausalLM"], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 4096, "intermediate_size": 11008, "initializer_range": 0.02, "max_sequence_length": 2048, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "pad_token_id": 0, "rms_norm_eps": 1e-06, "torch_dtype": "float16", "transformers_version": "4.27.0.dev0", "use_cache": true, "vocab_size": 32000}

generation_config.json

@@ -1 +1 @@
-{"_from_model_config": true, "bos_token_id": 0, "eos_token_id": 1, "pad_token_id": -1, "transformers_version": "4.27.0.dev0"}
+{"_from_model_config": true, "bos_token_id": 1, "eos_token_id": 2, "pad_token_id": 0, "transformers_version": "4.27.0.dev0"}

pytorch_model-00001-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:67b12ce16ab5dc0baab99b3f6e31b9dbae7d2389a5e67cfcf00538296cb981b0
+oid sha256:c5cb2ef8a6558c0ce314bcc11e4a15d0df379222ad7753ae9f75591cc87c945d
 size 404770755

pytorch_model-00002-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6c85ad85077dc5b9146212f9bff5752dcfc0198915f6004864c81d9901fa6367
+oid sha256:bba688e4ab77a5a95aebf52e0aa508d31ec1c02b6b5078191379c158526e6c12
 size 404770755

pytorch_model-00003-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:56c201febbcfdb9597d2541f0b8b6cf1009d9b582f777df3ec6b1846bbca26c5
+oid sha256:98ce0f7e5e8c374e697ded9576dfc5023f83e3c4c8a65dce63c009401672d2ca
 size 404770755

pytorch_model-00004-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b29426b2ebec2192577dbd3f2f1c478e876d9312ac5b88ea7fa2e206fb998a2c
+oid sha256:3cfe1d1093cbbb04d36b4fd09978e87077e41b4492b8cbba6bfff57c8dbd86dd
 size 404770755

pytorch_model-00005-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4790863e3883805020c03169f421c410dc319de3dd5ae19f6ff324e7b22327f1
+oid sha256:cc8014666b4da29ee637d96030f5e7612d2f2dff79a4df8ad4aafcb0125357d3
 size 404770755

pytorch_model-00006-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c8e31598150f786b03e922d68ac1622285bf703fb78fb1860829f3d8d72f23b7
+oid sha256:0d708f41d25ad5461d979337713757fabfe76c375b205b536bfe14d48556d5e0
 size 404770755

pytorch_model-00007-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:378b50bda5d8d07b18e67f4dbcd1132c4114a5819c0c906fc0306cfa815f9ed1
+oid sha256:9e6b78baebf20d60da825c3654c202008acfc3056c84cb6dadd91b155d278e38
 size 404770755

pytorch_model-00008-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb9f8d40af415d735f2c39133d95ce08dc67c3dc1980f42da64cffe068921300
+oid sha256:6080c736b3136935f0d3fb9fc65b41008c65b6b7d5238e1b9c4780043f31b955
 size 404770755

pytorch_model-00009-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe7477c5648709f0eeae1122ec01319bba538ba6d47bc973e0aee8cd68cd37a2
+oid sha256:90047b0adfef1abc939032be63cbe62300d3ee18b39251c2f36d362d6735b64e
 size 404770755

pytorch_model-00010-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52233530b595c9af47e31533a5a0e9bfb87c1d4ca9227268957ac637db6ff08d
+oid sha256:ec40a890dc20ddf9b49a6e1b18e26a304adbe701374a92f0295a7975139fa4b3
 size 404770755

pytorch_model-00011-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:028e101e316e45fb49b4b8eafa35b56f1e7b4fe35ae64c1a966bd72545aad020
+oid sha256:8d551e0c80f82a61a1cbf74389651359c3fb35ce50c4568256ebc8dcf7c1fcd3
 size 404770755

pytorch_model-00012-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e4eaa4637296f8f922e7c8bf489a315f375385973fd44db2e07f82d4a024bf4c
+oid sha256:869ac6cceca6c9f26d13ac05549b5e15d79acb90a0593475260171c29f3b9c73
 size 404770755

pytorch_model-00013-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9e87f3108c44616a18d5c995cf96b383b2cb4dc55ef516413832ce3dfd99ba88
+oid sha256:a18845375f10dc0085adc387f21e6ddc372338e767d1879c6369c11b27cb9b97
 size 404770755

pytorch_model-00014-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76b68d42b24321327b4a8f721083e1f1ff22ca0da810d017d4ce62b261171e1f
+oid sha256:c8e5dbe9332ea27857a2f9b8ef80e34567ff5d6f5855daffd73292080aedb401
 size 404770755

pytorch_model-00015-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:551845b71b470ebceddd62c993ed58f34a8e57987a8ae674619ce8bad462f4ec
+oid sha256:c259f397f43ccb73919b40810c4b33638bc655301a80805374800c124216ec80
 size 404770755

pytorch_model-00016-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:191f3c1e1b73ca2458630f54e2120829935d811de4f217f80c30240bce34e2e0
+oid sha256:ada30914f430170e601b4638ade39cbebb2cab280763ad5310492a35fed05082
 size 404770755

pytorch_model-00017-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f37b04b90a9535c71e2d7a7cf4baaae472c04adb343f7aa8b80f59e51d37443d
+oid sha256:f30a61cf2d7bef7ea0b927e05cd0f43d0aee6e00de73a9f7ba56f99deb4e3905
 size 404770755

pytorch_model-00018-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ab26083798343759b4915e4dd74cc3db92d1ad11890ae2180458165e87d9b081
+oid sha256:9e13fe2d35345fe65a1f7d4af98cf9a40f758f2403bd68a7e8ca4dbf50598014
 size 404770755

pytorch_model-00019-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fd02651e71dac125c30652db448f0c890e195a87683ee66b0d0ef5d568baea8f
+oid sha256:c993cd7044316208e3b8d8564d929e95bb57328dab63848fbb729b10049861eb
 size 404770755

pytorch_model-00020-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:031f663e3c9294d8c8f171570b25cfbc69852243a1e15a49ae26a6eca03bc355
+oid sha256:cb454682956e64a8c7c035764a6110dbb76fd7c61285160b42b06e7f0b5db0f2
 size 404770755

pytorch_model-00021-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:40587baf2075fb18757ab18e0254e096572adc7b977d4270071637c34e421b98
+oid sha256:1b35993de04e54aad4a034da33ae2860220482c0be5f5c2b9555c13f969f6531
 size 404770755

pytorch_model-00022-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ca63fa419f9a43aeff2e08e239ac7c6aa65b72226016cd36520e76d00c663cf5
+oid sha256:c3cb8f788c63a46e026ae54fca16aeb57d0af45e22bed01124dee12c266cf00d
 size 404770755

pytorch_model-00023-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fe9a235b1cce4432339fb25f6b02508d7ddce3432550ac04749444f9af9c0a0c
+oid sha256:cd034aad548c5423a8b4e07a4c44d8c2be1c350f9e8681cfdf0e63328fabadf9
 size 404770755

pytorch_model-00024-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:293f51504454408e8636c34fb40b29a3ad4e586dba6804e59f26e010dc10d834
+oid sha256:33d306be39a567f349b9e7bdc8b2543158607ca8834d8c71d32e1b648367ca94
 size 404770755

pytorch_model-00025-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1bd8aab9450cbc73ec907a3d0c720f7c73020fd03e0271ee4426e610554efff1
+oid sha256:d794d97063fb3c24ce913efa473b393dc6330aa67b00a47c4a91428de03896bc
 size 404770755

pytorch_model-00026-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f73d2e7f8b5330f5d6c051a77b3bfd1e7e1cbf387e78935684cac52e34bf2cce
+oid sha256:2aef4f0698b0670d6e9c83ca103c3c8dff70e2f237065e2c7293f496933ead0e
 size 404770755

pytorch_model-00027-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8f9bb10d65d9872da3979a3a5451aecee61846884864a01df2f676557cdf8e43
+oid sha256:9c06dbeda998897337596e8d98d53be174083a09a049640bf4f31eef37f56ea0
 size 404770755

pytorch_model-00028-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:02c00c239fd2de2fff25cadc3be1a52f8bbfd4e95171e73b6a8ccc3216d73c03
+oid sha256:d6b38acac958aadb83ae22e7317a82ff64caccde6f0c02539bd499e8777218af
 size 404770755

pytorch_model-00029-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e56b1dc46538a928d9ae56970df8afeb03497d1eeec3bae2a20baa75bbc73c2c
+oid sha256:798176e2a686e6ae5e9532d1943baaee618190f81f52311610c91603c6408fa7
 size 404770755

pytorch_model-00030-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e444f1d25002fc939434d48eb9ab7ff3d8898ed5b311936c3d9e2d0f5ebaf257
+oid sha256:ce7a9ef1002e65ff614d8237d6a643b4b6b96ef68b260532f219c692884f9b70
 size 404770755

pytorch_model-00031-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0552c5f16aa370cbfba24699b1b09d2246b42069d21099156ee5c25ed74eac9e
+oid sha256:3e746351f91d9d482c534620a441a7a66474fa9da57104b65f821fc513bad663
 size 404770755

pytorch_model-00032-of-00033.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:32cddee9159a708d3fb34735a0740a008a5867e6454456425133a4e65db135f0
-size 524297676
+oid sha256:95a865bd2b40a83e2110bb0a7de97dd08d7852b3858538ec5d65f64084667e44
+size 404770755

pytorch_model-00033-of-00033.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f289012ffe06b1aee0bebdfaea70a394c60444370d9b31c9ab0a7ccd7a4ee415
+size 524297676

pytorch_model.bin.index.json

@@ -1 +1 @@
-{"weight_map": {"model.layers.0.self_attn.q_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.mlp.down_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.mlp.up_proj.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.input_layernorm.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00000-of-00033.bin", "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00000-of-00033.bin", "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00033.bin", "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00033.bin", "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.mlp.down_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.mlp.up_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.input_layernorm.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00002-of-00033.bin", "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00033.bin", "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.mlp.down_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.mlp.up_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.input_layernorm.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00003-of-00033.bin", "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00033.bin", "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.mlp.down_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.mlp.up_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.input_layernorm.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00004-of-00033.bin", "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00033.bin", "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.mlp.down_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.mlp.up_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.input_layernorm.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00005-of-00033.bin", "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00033.bin", "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.mlp.down_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.mlp.up_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.input_layernorm.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00006-of-00033.bin", "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00033.bin", "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.mlp.down_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.mlp.up_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.input_layernorm.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00007-of-00033.bin", "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00033.bin", "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.mlp.down_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.mlp.up_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.input_layernorm.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00008-of-00033.bin", "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00033.bin", "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.mlp.down_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.mlp.up_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.input_layernorm.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00009-of-00033.bin", "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00033.bin", "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.mlp.down_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.mlp.up_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.input_layernorm.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00010-of-00033.bin", "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00033.bin", "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.mlp.down_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.mlp.up_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.input_layernorm.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00011-of-00033.bin", "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00033.bin", "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.mlp.down_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.mlp.up_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.input_layernorm.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00012-of-00033.bin", "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00033.bin", "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.mlp.down_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.mlp.up_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.input_layernorm.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00013-of-00033.bin", "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00033.bin", "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.mlp.down_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.mlp.up_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.input_layernorm.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00014-of-00033.bin", "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00033.bin", "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.mlp.down_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.mlp.up_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.input_layernorm.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00015-of-00033.bin", "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00033.bin", "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.mlp.down_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.mlp.up_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.input_layernorm.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00016-of-00033.bin", "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00033.bin", "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.mlp.down_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.mlp.up_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.input_layernorm.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00017-of-00033.bin", "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00033.bin", "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.mlp.down_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.mlp.up_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.input_layernorm.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00018-of-00033.bin", "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00033.bin", "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.mlp.down_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.mlp.up_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.input_layernorm.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00019-of-00033.bin", "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00033.bin", "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.mlp.down_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.mlp.up_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.input_layernorm.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00020-of-00033.bin", "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00033.bin", "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.mlp.down_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.mlp.up_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.input_layernorm.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00021-of-00033.bin", "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00033.bin", "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.mlp.down_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.mlp.up_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.input_layernorm.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00022-of-00033.bin", "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00033.bin", "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.mlp.down_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.mlp.up_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.input_layernorm.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00023-of-00033.bin", "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00033.bin", "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.mlp.down_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.mlp.up_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.input_layernorm.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00024-of-00033.bin", "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00033.bin", "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.mlp.down_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.mlp.up_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.input_layernorm.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00025-of-00033.bin", "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00033.bin", "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.mlp.down_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.mlp.up_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.input_layernorm.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00026-of-00033.bin", "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00033.bin", "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.mlp.down_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.mlp.up_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.input_layernorm.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00027-of-00033.bin", "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00033.bin", "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.mlp.down_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.mlp.up_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.input_layernorm.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00028-of-00033.bin", "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00033.bin", "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.mlp.down_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.mlp.up_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.input_layernorm.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00029-of-00033.bin", "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00033.bin", "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.mlp.down_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.mlp.up_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.input_layernorm.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00030-of-00033.bin", "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00033.bin", "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.mlp.down_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.mlp.up_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.input_layernorm.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00031-of-00033.bin", "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00033.bin", "model.embed_tokens.weight": "pytorch_model-00032-of-00033.bin", "model.norm.weight": "pytorch_model-00032-of-00033.bin", "lm_head.weight": "pytorch_model-00032-of-00033.bin"}, "metadata": {"total_size": 13476835328}}
+{"weight_map": {"model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00033.bin", "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00033.bin", "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.mlp.down_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.mlp.up_proj.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.input_layernorm.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00002-of-00033.bin", "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00033.bin", "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.mlp.down_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.mlp.up_proj.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.input_layernorm.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00003-of-00033.bin", "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00033.bin", "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.mlp.down_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.mlp.up_proj.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.input_layernorm.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00004-of-00033.bin", "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00033.bin", "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.mlp.down_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.mlp.up_proj.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.input_layernorm.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00005-of-00033.bin", "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00033.bin", "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.mlp.down_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.mlp.up_proj.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.input_layernorm.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00006-of-00033.bin", "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00033.bin", "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.mlp.down_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.mlp.up_proj.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.input_layernorm.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00007-of-00033.bin", "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00033.bin", "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.mlp.down_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.mlp.up_proj.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.input_layernorm.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00008-of-00033.bin", "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00033.bin", "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.mlp.down_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.mlp.up_proj.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.input_layernorm.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00009-of-00033.bin", "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00033.bin", "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.mlp.down_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.mlp.up_proj.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.input_layernorm.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00010-of-00033.bin", "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00033.bin", "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.mlp.down_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.mlp.up_proj.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.input_layernorm.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00011-of-00033.bin", "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00033.bin", "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.mlp.down_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.mlp.up_proj.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.input_layernorm.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00012-of-00033.bin", "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00033.bin", "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.mlp.down_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.mlp.up_proj.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.input_layernorm.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00013-of-00033.bin", "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00033.bin", "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.mlp.down_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.mlp.up_proj.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.input_layernorm.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00014-of-00033.bin", "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00033.bin", "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.mlp.down_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.mlp.up_proj.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.input_layernorm.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00015-of-00033.bin", "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00033.bin", "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.mlp.down_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.mlp.up_proj.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.input_layernorm.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00016-of-00033.bin", "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00033.bin", "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.mlp.down_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.mlp.up_proj.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.input_layernorm.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00017-of-00033.bin", "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00033.bin", "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.mlp.down_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.mlp.up_proj.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.input_layernorm.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00018-of-00033.bin", "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00033.bin", "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.mlp.down_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.mlp.up_proj.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.input_layernorm.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00019-of-00033.bin", "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00033.bin", "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.mlp.down_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.mlp.up_proj.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.input_layernorm.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00020-of-00033.bin", "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00033.bin", "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.mlp.down_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.mlp.up_proj.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.input_layernorm.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00021-of-00033.bin", "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00033.bin", "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.mlp.down_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.mlp.up_proj.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.input_layernorm.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00022-of-00033.bin", "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00033.bin", "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.mlp.down_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.mlp.up_proj.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.input_layernorm.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00023-of-00033.bin", "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00033.bin", "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.mlp.down_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.mlp.up_proj.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.input_layernorm.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00024-of-00033.bin", "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00033.bin", "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.mlp.down_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.mlp.up_proj.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.input_layernorm.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00025-of-00033.bin", "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00033.bin", "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.mlp.down_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.mlp.up_proj.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.input_layernorm.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00026-of-00033.bin", "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00033.bin", "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.mlp.down_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.mlp.up_proj.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.input_layernorm.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00027-of-00033.bin", "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00033.bin", "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.mlp.down_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.mlp.up_proj.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.input_layernorm.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00028-of-00033.bin", "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00033.bin", "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.mlp.down_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.mlp.up_proj.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.input_layernorm.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00029-of-00033.bin", "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00033.bin", "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.mlp.down_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.mlp.up_proj.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.input_layernorm.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00030-of-00033.bin", "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00033.bin", "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.mlp.down_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.mlp.up_proj.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.input_layernorm.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00031-of-00033.bin", "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00033.bin", "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.mlp.down_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.mlp.up_proj.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.input_layernorm.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00032-of-00033.bin", "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00033.bin", "model.embed_tokens.weight": "pytorch_model-00033-of-00033.bin", "model.norm.weight": "pytorch_model-00033-of-00033.bin", "lm_head.weight": "pytorch_model-00033-of-00033.bin"}, "metadata": {"total_size": 13476835328}}

tokenizer_config.json

@@ -1 +1 @@
-{"bos_token": "", "eos_token": "", "model_max_length": 1000000000000000019884624838656, "tokenizer_class": "LLaMATokenizer", "unk_token": ""}
+{"bos_token": "", "eos_token": "", "model_max_length": 1000000000000000019884624838656, "tokenizer_class": "LlamaTokenizer", "unk_token": ""}