riczhou commited on
Commit
4a37ca9
·
verified ·
1 Parent(s): 765d853

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
added_tokens.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "<image_soft_token>": 262144
3
+ }
mlc-chat-config.json ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.1.0",
3
+ "model_type": "gemma3_text",
4
+ "quantization": "q0bf16",
5
+ "model_config": {
6
+ "text_config": {
7
+ "hidden_size": 1152,
8
+ "intermediate_size": 6912,
9
+ "num_hidden_layers": 26,
10
+ "attention_bias": false,
11
+ "num_attention_heads": 4,
12
+ "num_key_value_heads": 1,
13
+ "head_dim": 256,
14
+ "rms_norm_eps": 1e-06,
15
+ "hidden_activation": "gelu_pytorch_tanh",
16
+ "position_embedding_base": 1000000,
17
+ "context_window_size": 8192,
18
+ "prefill_chunk_size": 8192,
19
+ "query_pre_attn_scalar": 256,
20
+ "sliding_window": 512,
21
+ "kwargs": {
22
+ "architectures": [
23
+ "Gemma3ForCausalLM"
24
+ ],
25
+ "attention_dropout": 0.0,
26
+ "attn_logit_softcapping": null,
27
+ "bos_token_id": 2,
28
+ "cache_implementation": "hybrid",
29
+ "eos_token_id": [
30
+ 1,
31
+ 106
32
+ ],
33
+ "final_logit_softcapping": null,
34
+ "initializer_range": 0.02,
35
+ "max_position_embeddings": 32768,
36
+ "model_type": "gemma3_text",
37
+ "pad_token_id": 0,
38
+ "rope_local_base_freq": 10000,
39
+ "rope_scaling": null,
40
+ "sliding_window_pattern": 6,
41
+ "torch_dtype": "bfloat16",
42
+ "transformers_version": "4.50.0.dev0",
43
+ "use_cache": true
44
+ }
45
+ },
46
+ "vocab_size": 262144,
47
+ "tensor_parallel_shards": 1,
48
+ "max_batch_size": 128,
49
+ "context_window_size": 8192,
50
+ "sliding_window_size": -1,
51
+ "prefill_chunk_size": 8192,
52
+ "is_text_model": true
53
+ },
54
+ "vocab_size": 262144,
55
+ "context_window_size": 8192,
56
+ "sliding_window_size": -1,
57
+ "prefill_chunk_size": 8192,
58
+ "attention_sink_size": -1,
59
+ "tensor_parallel_shards": 1,
60
+ "pipeline_parallel_stages": 1,
61
+ "temperature": 1.0,
62
+ "presence_penalty": 0.0,
63
+ "frequency_penalty": 0.0,
64
+ "repetition_penalty": 1.0,
65
+ "top_p": 1.0,
66
+ "tokenizer_files": [
67
+ "tokenizer.model",
68
+ "tokenizer.json",
69
+ "added_tokens.json",
70
+ "tokenizer_config.json"
71
+ ],
72
+ "tokenizer_info": {
73
+ "token_postproc_method": "byte_fallback",
74
+ "prepend_space_in_encode": false,
75
+ "strip_space_in_decode": false
76
+ },
77
+ "conv_template": {
78
+ "name": "gemma_instruction",
79
+ "system_template": "{system_message}",
80
+ "system_message": "",
81
+ "system_prefix_token_ids": [
82
+ 2
83
+ ],
84
+ "add_role_after_system_message": true,
85
+ "roles": {
86
+ "user": "<start_of_turn>user",
87
+ "assistant": "<start_of_turn>model"
88
+ },
89
+ "role_templates": {
90
+ "user": "{user_message}",
91
+ "assistant": "{assistant_message}",
92
+ "tool": "{tool_message}"
93
+ },
94
+ "messages": [],
95
+ "seps": [
96
+ "<end_of_turn>\n"
97
+ ],
98
+ "role_content_sep": "\n",
99
+ "role_empty_sep": "\n",
100
+ "stop_str": [
101
+ "<end_of_turn>"
102
+ ],
103
+ "stop_token_ids": [
104
+ 1,
105
+ 107
106
+ ],
107
+ "function_string": "",
108
+ "use_function_calling": false
109
+ },
110
+ "pad_token_id": 0,
111
+ "bos_token_id": 2,
112
+ "eos_token_id": [
113
+ 1,
114
+ 106
115
+ ]
116
+ }
ndarray-cache.json ADDED
The diff for this file is too large to render. See raw diff
 
params_shard_0.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8169474a8e8d2df7cc2ab19a5dd6aed0c2fe51afa2d9bfff2fc5235eea2bae26
3
+ size 603979776
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2595b11b9a7cf27e9c4e8fe5f400cb99fca602179204fa2093799bb55447958b
3
+ size 31850496
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52945825bba5c037700fa8751a2b53507d46997a82f0d968de48636a27662c5a
3
+ size 21833728
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:630208660071fddd409a77a8452fea0025b56d47e88c367df10b0501b91c99d5
3
+ size 31850496
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51980fc502f9622d773d0a84a47a2f866bb9035229a9f9d49c060732d0844e07
3
+ size 21833728
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ca34d9f2dd673de0aa59ea98df20222bc8723ede2c7841cee98af903d79c2c9
3
+ size 31850496
params_shard_14.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f04e5eddddbbcd82466249c2dd5c46e01df023a251c7f5b35e1e574ba9041249
3
+ size 21833728
params_shard_15.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45d4fd47e61adf8c3bd1be047131cc19d53b503af5e69db0804f8a3e9308fbf2
3
+ size 31850496
params_shard_16.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f1e385e9501036330de616f19a35ea200ad3d5331829be855bf814a50fa1dee
3
+ size 21833728
params_shard_17.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1253e760de9d9ba6f50dc1b127641f7089e6cb0653dd7fafc577a47df74c6b9
3
+ size 31850496
params_shard_18.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9cdcb3f2cf1c004aefb3d3f1214f82299723c310c6ffab276bdae6fde11595f4
3
+ size 21833728
params_shard_19.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37731793d5c65eee110ad99aa8ed7204fd0441c8569f981798cfa88c6a335bea
3
+ size 31850496
params_shard_2.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:29b02851cf25fde0e0a4d6316740f8bc14da50d45a776a7ce0ff57eb975f3a1f
3
+ size 21836032
params_shard_20.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78e4fe0ddda0ff817e371502d0ab7d96b355df42fc744de90640304bddec8b4a
3
+ size 21833728
params_shard_21.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8630ed2832d7311f8ff31bc56c3400973e0044c68b5e36c67c5959868d9b5051
3
+ size 31850496
params_shard_22.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c34c656caca8c7fba932681d8a126f90639b6677bbce16ebf1afc4aebb53644
3
+ size 21833728
params_shard_23.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:105e2f3ff9e7ff22603ac26e2d75a66c1b661d3d9ed765e8bd2d7e6ef36c0d81
3
+ size 31850496
params_shard_24.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dcdfc6cfb1acc05b08c9fefbc7d1c82a2d9f574b51736a392c0c8b844c4fcc68
3
+ size 21833728
params_shard_25.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fab3685c852cfb0b78b3135896a5f7dc3a7ba37567a6b4d992aaa03a76c65618
3
+ size 31850496
params_shard_26.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d207ba700331ba1fc834272199911c7c69eb52497cf77cc7af82bdd689135ace
3
+ size 21833728
params_shard_27.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8ae1de90d0b7dfe29bfb74fa8da367851648fb8a7460f0d6c27bb029ef6eedd
3
+ size 31850496
params_shard_28.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f23a0ea64b922b4b46aa7030fdd71bee736c8cfef8ab91ec5048372769f59534
3
+ size 21833728
params_shard_29.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3289f3131c071eb9381e32d00b792444de1198f258caaba659e6ba056bee446f
3
+ size 31850496
params_shard_3.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5ef6482078589e2719bf7dbe96d33d51db68503259e58cf27c52fd0e1ceb5ec
3
+ size 31850496
params_shard_30.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7185673606422fa8d4dc0f6253aabc0b8e76dfd61b8b5237d6b015c1b18264a3
3
+ size 21833728
params_shard_31.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4cea6a928748dbcbdb27fd485bf94646f413fae13a30aabf85988588353f637
3
+ size 31850496
params_shard_32.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad0445e679c798e517666a2f2de7491793268f538df8f4cd4e77d233d10cf0ea
3
+ size 21833728
params_shard_33.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2faadec2c4a6745e57e1e467f6c98424bf2a36b746da8741d6a5712d7eb45015
3
+ size 31850496
params_shard_34.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d4dbca031c873d3c29a9e5adc406694322c1cbbf89417f3e841d63b3e69ae58
3
+ size 21833728
params_shard_35.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d030eb855a8dc99848bc948bc83bd8dd8bfe6f3e11af06608fde4b0448ba3055
3
+ size 31850496
params_shard_36.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f39d66f06d93a418cc82dc4759f8e2fec179b65081eba52e72b8d302c12b0f06
3
+ size 21833728
params_shard_37.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:626868228097d711c74e7b38386b8308ea9b11e735aa8647fb057d72b6a1741f
3
+ size 31850496
params_shard_38.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:32457987ab4247a337ef9ade41a2e069cfd2abec3bf003ce3220031501ad4c25
3
+ size 21833728
params_shard_39.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0dddffe70ff25e1628082c71be92b46c47bc4a3b6e6c01c0caa851999729a181
3
+ size 31850496
params_shard_4.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e66de86a49f67bc543b4df464c40d2ae1eeb958cc702b3e9efa03d8f67319801
3
+ size 21833728
params_shard_40.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ed41d9be443d7ea20e9b05ca5dcb484420be0c0d2ebdaebb4fcc61148345233
3
+ size 21833728
params_shard_41.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2e2cabf7760c51ff9c38f4a752c34bd0af8a197cba9f71761596f4578a29185
3
+ size 31850496
params_shard_42.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26a26897aecf8a67edfcb4fa0226bf241b0ac5296abb9a78dae730cf0b9298d4
3
+ size 21833728
params_shard_43.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:34a47bedb9c324c3f23720ecb9e52bc95282a4591b97e94cb3e75d0f2debdd40
3
+ size 31850496
params_shard_44.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ada976190640cfef513901e33c9a7cd784ca329c2197b3f1c3f2def68258caa9
3
+ size 21833728
params_shard_45.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7579689758501ce129e322fa36ab5868a7684ec33df089728ee557ddce9b8818
3
+ size 31850496
params_shard_46.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80e716760dc833457f7a9b8f2e865b59d4c917ca7f4eb8b0d145f7a34ca520d5
3
+ size 21833728
params_shard_47.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c0055ba9f5258bf2abcc06199fc487b432e031747288e898cb38e88fc0414df
3
+ size 31850496
params_shard_48.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22f47d0d2c8570413b6bc10ec6f364b307908ce038842a2b9f02bc65444fc46f
3
+ size 21833728
params_shard_49.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7dbc9982e57cd723250c62be85af68298e676ab4a362c283ec3b1fd5e6b9990
3
+ size 31850496
params_shard_5.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:30807a7fad3043d391b44c36d951be847fb8ef1972cf7b0e128b458d69764057
3
+ size 31850496