riczhou commited on
Commit
039a0fc
·
verified ·
1 Parent(s): fb8d7cc

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
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
mlc-chat-config.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "0.1.0",
3
+ "model_type": "qwen3_moe",
4
+ "quantization": "q4f32_1",
5
+ "model_config": {
6
+ "hidden_act": "silu",
7
+ "hidden_size": 2048,
8
+ "intermediate_size": 6144,
9
+ "attention_bias": false,
10
+ "num_attention_heads": 32,
11
+ "num_hidden_layers": 48,
12
+ "num_key_value_heads": 4,
13
+ "rms_norm_eps": 1e-06,
14
+ "rope_theta": 1000000.0,
15
+ "vocab_size": 151936,
16
+ "tie_word_embeddings": false,
17
+ "context_window_size": 40960,
18
+ "prefill_chunk_size": 2048,
19
+ "tensor_parallel_shards": 1,
20
+ "head_dim": 128,
21
+ "dtype": "float32",
22
+ "max_batch_size": 128,
23
+ "moe_intermediate_size": 768,
24
+ "num_experts_per_tok": 8,
25
+ "num_experts": 128,
26
+ "decoder_sparse_step": 1,
27
+ "norm_topk_prob": true
28
+ },
29
+ "vocab_size": 151936,
30
+ "context_window_size": 40960,
31
+ "sliding_window_size": -1,
32
+ "prefill_chunk_size": 2048,
33
+ "attention_sink_size": -1,
34
+ "tensor_parallel_shards": 1,
35
+ "pipeline_parallel_stages": 1,
36
+ "temperature": 0.6,
37
+ "presence_penalty": 0.0,
38
+ "frequency_penalty": 0.0,
39
+ "repetition_penalty": 1.0,
40
+ "top_p": 0.95,
41
+ "tokenizer_files": [
42
+ "tokenizer.json",
43
+ "vocab.json",
44
+ "merges.txt",
45
+ "tokenizer_config.json"
46
+ ],
47
+ "tokenizer_info": {
48
+ "token_postproc_method": "byte_level",
49
+ "prepend_space_in_encode": false,
50
+ "strip_space_in_decode": false
51
+ },
52
+ "conv_template": {
53
+ "name": "qwen2",
54
+ "system_template": "<|im_start|>system\n{system_message}<|im_end|>\n",
55
+ "system_message": "You are a helpful assistant.",
56
+ "system_prefix_token_ids": null,
57
+ "add_role_after_system_message": true,
58
+ "roles": {
59
+ "user": "<|im_start|>user",
60
+ "assistant": "<|im_start|>assistant"
61
+ },
62
+ "role_templates": {
63
+ "user": "{user_message}",
64
+ "assistant": "{assistant_message}",
65
+ "tool": "{tool_message}"
66
+ },
67
+ "messages": [],
68
+ "seps": [
69
+ "<|im_end|>\n"
70
+ ],
71
+ "role_content_sep": "\n",
72
+ "role_empty_sep": "\n",
73
+ "stop_str": [
74
+ "<|endoftext|>",
75
+ "<|im_end|>"
76
+ ],
77
+ "stop_token_ids": [
78
+ 151643,
79
+ 151645
80
+ ],
81
+ "function_string": "",
82
+ "use_function_calling": false
83
+ },
84
+ "pad_token_id": 151643,
85
+ "bos_token_id": 151643,
86
+ "eos_token_id": [
87
+ 151645,
88
+ 151643
89
+ ]
90
+ }
ndarray-cache-b16.json ADDED
The diff for this file is too large to render. See raw diff
 
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 sha256:47358095200b1d055d34722c0c051e438c1e39a6c679f18b3eb0b833ce49872e
3
+ size 155582464
params_shard_1.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad02d667e6760c97e3cdf08e4a99a6f89c2cba81af734fe5cc87d50d001d8acf
3
+ size 100663296
params_shard_10.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e51fdd1490e6f8a650d8bc3a5b78754e67f3900542384d3ef27e2e9f0f0bc0e
3
+ size 100663296
params_shard_100.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65e94a048cbdef082214d7fe4f2873ac1b584a37f2c5f0242df7326f362d82e9
3
+ size 25165824
params_shard_101.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af58376d549def8b7182e8e7526ff0fc3a8ccc255372da5720cc8b1ea0d904f3
3
+ size 100663296
params_shard_102.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5ed8b7ce6bed0f98fab0c6d80a875ead2da26592a01b24d95dd0fd2f9e030a5
3
+ size 23732736
params_shard_103.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20313318d830f1f621abd7f1c077b94688536048dbc730aaa9a793c1087555b0
3
+ size 201326592
params_shard_104.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc84a13f0fe33c672841cb996b0ffd650b2442d40fdab7bc8846b848a9435362
3
+ size 25165824
params_shard_105.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eef528b07d9a3e8a11c55fadba135514a2caec9d3bec6bb6cfef1555a7010d65
3
+ size 100663296
params_shard_106.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5df91f49e303d8af456a82e122bbdf3d622aedf43c6b00d3a4423ec17322a67
3
+ size 23728640
params_shard_107.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:504439f0db6e6e3bc8f9769cb9fc02c9379efe4e8000bd02509c8ea62cc15648
3
+ size 201326592
params_shard_108.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2aeabef2b2971add8b3cce5c6ef54f658cf491d5288bde3c03606ee8adf5d1bf
3
+ size 25165824
params_shard_109.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bf6f93c60722ec73cf50dacbb147162e74100659d37bb206739c90bab73f470
3
+ size 100663296
params_shard_11.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f6ac258e9cb25150d5f314c8b36b77b8ab6429d8eea7626f4d5f4a77f355b11
3
+ size 23732736
params_shard_110.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f9958c00772cd5e1b8ae40c623d5a41453613e254f1e4e4ddb884ff34630d1a
3
+ size 23736832
params_shard_111.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:903bb2ecefafb377bdeef99836b3c99571eb5dcd222871df92d3488f93afa50a
3
+ size 201326592
params_shard_112.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3efd8b14135b164cb2df7e9c737b268c0940da7bfbcdfb51b840fa0757825afc
3
+ size 25165824
params_shard_113.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d6f54b86c0786a3445778b8a9318668275fb33f8997726fa6e0d5f395b787b8
3
+ size 100663296
params_shard_114.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2603ea34043f0094188922d09a3c4a71ced93c7ea26181c797053e984c8d80f1
3
+ size 23732736
params_shard_115.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f415cfc310a74990be0c5d51ed28bfd52c5152a107ce39fd3f7b45314709c664
3
+ size 201326592
params_shard_116.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4639be79f22656082b6d20816700349bea0bfdffdc2f1a5d2a35184ea0313945
3
+ size 25165824
params_shard_117.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b7d77cf56a3ea42da95268238009c603419a9be0264849b416b0fd7be5047d4
3
+ size 100663296
params_shard_118.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bdf44452a0c35bad25d8fef276e237df8e49a98006250587c5e0cc2abe2b0da5
3
+ size 23728640
params_shard_119.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5d485287d9e46cfbfbcd9f71c32d483218fa178bc9bb90c2fa4670134fce6f7
3
+ size 201326592
params_shard_12.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e569da2d1bf3a4814bd37b48f2cff285e804bad1284b8aa67260b591f0e8b46
3
+ size 201326592
params_shard_120.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0bd67cfafd97030af53744761d25a23d9762771a272bd75039db0ce9d9131427
3
+ size 25165824
params_shard_121.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f3174b7a141028a488e3b0053adee4068801d0967b0cb2fe0469a86a6b354917
3
+ size 100663296
params_shard_122.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e45814ec775db949e2fd5c61730e625f276190cb8651585abb5cabe82e377d75
3
+ size 23736832
params_shard_123.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bc14c6cc840feb5ffa4b9f66a783e011d44a9f47f1f06b99ad4b4e6e612e9ee
3
+ size 201326592
params_shard_124.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37796e79139b39f071b5ef7422bc2c146beedb9aa26004455ab946d08d2e6c20
3
+ size 25165824
params_shard_125.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bbac23663eeca2abf50f9cf119225239af17a88d911a8e6471b45c455d5638ee
3
+ size 100663296
params_shard_126.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04f8f00d83703e4c29e6a29326a3e57900fca8d72f95de28cefaebd6d363ebc1
3
+ size 23732736
params_shard_127.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:747265ed84b1d7febdaf371449fcfbbfd8b4c2ef17da281861676f22e341c453
3
+ size 201326592
params_shard_128.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5d482fb2ab2a45f9f72e924eafaad5c097db870760709d4a4c909c2678f5e0e
3
+ size 25165824
params_shard_129.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0527e1b0ba1503e117211e60a6316d4d6d7012023d5d899167efd86733994745
3
+ size 100663296
params_shard_13.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8255cb4a02028097709ac8b3f9c1f90b5a00c2db17a607852650b5001b3e108e
3
+ size 25165824
params_shard_130.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df3bb0892fe5e7254dcf8b78cc91ab69bf75a47044ccf8d3a9e06763a0a62763
3
+ size 23728640
params_shard_131.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8967f94cbd8b0bfe99c4c7a9584ff2d7c71cfa1659b09373b9e867a23c56b009
3
+ size 201326592
params_shard_132.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4110d9b2157696f771d70c7e8a4bc0a2460632e3868de806e10d150d672dde9
3
+ size 25165824
params_shard_133.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1a53465e65b11e910c96e5943fc6473226ec11c7b739f530aad894791684c22
3
+ size 100663296
params_shard_134.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39ebaa71100946715be8fbee48c4fe4abb0b8c17cde7344f63633deabdb40c09
3
+ size 23736832
params_shard_135.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7cc651637db2ccea5384fb9ae366d5d03819baf4243f6e8b3e4842ae2aec11c
3
+ size 201326592
params_shard_136.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3bb4237d0f14f9bb0c9a80263c6f4dfd7ee6263f92ae6e5c01d69ddb8bb864d
3
+ size 25165824
params_shard_137.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13f783e40d30fae2a1ac6fecbfb90220f191a1e1bfe7cdcf0e8e7acba3cad3f5
3
+ size 100663296
params_shard_138.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:29a9e839bd7f8640829ab2f3f176f8bcffec3cd76613e0227776aa6b0d9a4af9
3
+ size 23732736