kishizaki-sci
commited on
Upload LlamaForCausalLM
Browse files- config.json +40 -0
- generation_config.json +11 -0
- model-00001-of-00044.safetensors +3 -0
- model-00002-of-00044.safetensors +3 -0
- model-00003-of-00044.safetensors +3 -0
- model-00004-of-00044.safetensors +3 -0
- model-00005-of-00044.safetensors +3 -0
- model-00006-of-00044.safetensors +3 -0
- model-00007-of-00044.safetensors +3 -0
- model-00008-of-00044.safetensors +3 -0
- model-00009-of-00044.safetensors +3 -0
- model-00010-of-00044.safetensors +3 -0
- model-00011-of-00044.safetensors +3 -0
- model-00012-of-00044.safetensors +3 -0
- model-00013-of-00044.safetensors +3 -0
- model-00014-of-00044.safetensors +3 -0
- model-00015-of-00044.safetensors +3 -0
- model-00016-of-00044.safetensors +3 -0
- model-00017-of-00044.safetensors +3 -0
- model-00018-of-00044.safetensors +3 -0
- model-00019-of-00044.safetensors +3 -0
- model-00020-of-00044.safetensors +3 -0
- model-00021-of-00044.safetensors +3 -0
- model-00022-of-00044.safetensors +3 -0
- model-00023-of-00044.safetensors +3 -0
- model-00024-of-00044.safetensors +3 -0
- model-00025-of-00044.safetensors +3 -0
- model-00026-of-00044.safetensors +3 -0
- model-00027-of-00044.safetensors +3 -0
- model-00028-of-00044.safetensors +3 -0
- model-00029-of-00044.safetensors +3 -0
- model-00030-of-00044.safetensors +3 -0
- model-00031-of-00044.safetensors +3 -0
- model-00032-of-00044.safetensors +3 -0
- model-00033-of-00044.safetensors +3 -0
- model-00034-of-00044.safetensors +3 -0
- model-00035-of-00044.safetensors +3 -0
- model-00036-of-00044.safetensors +3 -0
- model-00037-of-00044.safetensors +3 -0
- model-00038-of-00044.safetensors +3 -0
- model-00039-of-00044.safetensors +3 -0
- model-00040-of-00044.safetensors +3 -0
- model-00041-of-00044.safetensors +3 -0
- model-00042-of-00044.safetensors +3 -0
- model-00043-of-00044.safetensors +3 -0
- model-00044-of-00044.safetensors +3 -0
- model.safetensors.index.json +0 -0
config.json
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "hf_cache/Llama-3.1-405B-Instruct-f16",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 128000,
|
9 |
+
"eos_token_id": [
|
10 |
+
128001,
|
11 |
+
128008,
|
12 |
+
128009
|
13 |
+
],
|
14 |
+
"head_dim": 128,
|
15 |
+
"hidden_act": "silu",
|
16 |
+
"hidden_size": 16384,
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 53248,
|
19 |
+
"max_position_embeddings": 131072,
|
20 |
+
"mlp_bias": false,
|
21 |
+
"model_type": "llama",
|
22 |
+
"num_attention_heads": 128,
|
23 |
+
"num_hidden_layers": 126,
|
24 |
+
"num_key_value_heads": 8,
|
25 |
+
"pretraining_tp": 1,
|
26 |
+
"rms_norm_eps": 1e-05,
|
27 |
+
"rope_scaling": {
|
28 |
+
"factor": 8.0,
|
29 |
+
"high_freq_factor": 4.0,
|
30 |
+
"low_freq_factor": 1.0,
|
31 |
+
"original_max_position_embeddings": 8192,
|
32 |
+
"rope_type": "llama3"
|
33 |
+
},
|
34 |
+
"rope_theta": 500000.0,
|
35 |
+
"tie_word_embeddings": false,
|
36 |
+
"torch_dtype": "float16",
|
37 |
+
"transformers_version": "4.47.0.dev0",
|
38 |
+
"use_cache": false,
|
39 |
+
"vocab_size": 128256
|
40 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 128000,
|
4 |
+
"eos_token_id": [
|
5 |
+
128001,
|
6 |
+
128008,
|
7 |
+
128009
|
8 |
+
],
|
9 |
+
"transformers_version": "4.47.0.dev0",
|
10 |
+
"use_cache": false
|
11 |
+
}
|
model-00001-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9bba25a57af94f30c5474eda6a3aa61c1a0398987afd85f311c1eb546350af54
|
3 |
+
size 4952295224
|
model-00002-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b99710ae69ed3341c35f6d90c796c0b5b8403c7b58251abedc273180b78c8a9a
|
3 |
+
size 4968488696
|
model-00003-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c7c4f6e17af4ca259ecab776dbd079756fbf8fcef78ce93f6cfc139e415e37d7
|
3 |
+
size 4968488696
|
model-00004-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0e126a9a1aa464c98e3576f29e07e9448fbef73f505a61f067bdcaa6954a3663
|
3 |
+
size 4968488696
|
model-00005-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:869fd9611d752c8b80a21078e38ed29d0ae20391cdaaeb4a3ab310d77f786f6e
|
3 |
+
size 4968488760
|
model-00006-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:459a15e41e851dac3be6fa2d446c95f5754038994cbc1929c836420ceeb98f97
|
3 |
+
size 4968488768
|
model-00007-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:eacaf42dd9fac79584e3f8029aceaaa16942d6ed152f40bf3aaf065d499bb100
|
3 |
+
size 4968488768
|
model-00008-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb7f5932e499aff5885fc050c8664c83f41ecc4c3022911868f998be0dc82dd0
|
3 |
+
size 4968488768
|
model-00009-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5143c9d64c880ed2f202d991e3a7c685fcc78238a1634b87a1c61f5cb9e9660f
|
3 |
+
size 4968488768
|
model-00010-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cecef1181d63271e25fda1bf7386f6ca2d73e2f8cf170c80713468adff0eddf9
|
3 |
+
size 4968488768
|
model-00011-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b9268e8e5796406a209599a1917cf9003154b4a92f38b58447c339e6acce7c0c
|
3 |
+
size 4968488768
|
model-00012-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9dd4ac8085f1e913376aae3a1f8f2f260a2b038ab692f6cf7931f0d5c3c63efd
|
3 |
+
size 4968488768
|
model-00013-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f4d90a77cb51209a1aff5ec2c45e47fe8f664f062ce24fb56e88de60b004bb3d
|
3 |
+
size 4968488768
|
model-00014-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fed7ace4783fb5acbd9e262ff03e04975f4ef5a22f92584b9505581e9b093c3c
|
3 |
+
size 4968488768
|
model-00015-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0e3ceae1da710f2318ab9d635653f4f24b36da0c6b5656562c84972b1e1ef318
|
3 |
+
size 4968488768
|
model-00016-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:34d4c642fcb14319733bac447a609e3743a84235edbf67ca16fcb1e7e46ed5df
|
3 |
+
size 4968488768
|
model-00017-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:985d612abf85b0eeed225edf8977bd190b5c4d45486d879257fc209bcc277582
|
3 |
+
size 4968488768
|
model-00018-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:33d3982fc5cc690f08d48b491cc1014fcc340d88b5578760ebc75eb421b1b02b
|
3 |
+
size 4968488768
|
model-00019-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:dc7b82b5aa13197a82eacff9db4c419b74c6b1f22b71136599aeb87924597044
|
3 |
+
size 4968488768
|
model-00020-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a99753cca9628ff8800af80c7010e15972a56b7085786f6a8ae6a2ba7fd2391
|
3 |
+
size 4968488768
|
model-00021-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5aa8d38a0e6899b9527bac46fbdd7b14f14944b69686b73223674ed859dbe744
|
3 |
+
size 4968488768
|
model-00022-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:edfe2e5a8c4d772df6fbe6d6a6f3fbe5fa02b3aa7769c4ab6e98f0db26da1554
|
3 |
+
size 4968488768
|
model-00023-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:32cdf1de42e011a5a061e25d0813ecd4a2c468d822de17a3b676074b7e9d5c27
|
3 |
+
size 4968488768
|
model-00024-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8cec2116829ca40c1bc7edf87a0359246c720b4d675eb9a6ddda04e330a94115
|
3 |
+
size 4968488768
|
model-00025-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fc93ff9afc24b59a75ed68a1714833ce7ff893af31dd13d016f133cb05ff557c
|
3 |
+
size 4968488768
|
model-00026-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8095b1a64b8711cb42e036b28e1fda6dc8018f36ae6947b88a927cc6d0591cd6
|
3 |
+
size 4968488768
|
model-00027-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:360b3c126ccd6507b62c7b679f261b00a82bfef986f43902b24d3b49bf5f53f5
|
3 |
+
size 4968488768
|
model-00028-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:aa9e45b720496129ff591f673b4d3290f5668e0a91507a23582112b802530b0d
|
3 |
+
size 4968488768
|
model-00029-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f3c1fb4569f8599bb7c423297a6c8cf38292f52a96d099b8e4487d9cd79b65a2
|
3 |
+
size 4968488768
|
model-00030-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:390b9ef2655b4eee36403197ee26626c9820dc99b3d6869f734f6bf923030ab5
|
3 |
+
size 4968488768
|
model-00031-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb8201640f189f290009b019eb1e32420bf25c3b86be33ed04e4eab4f86de817
|
3 |
+
size 4968488768
|
model-00032-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a4fad25043dbeb70d3e7b63077d81a62325ae9dc6436d57cdbc04e1ebcd1c8c1
|
3 |
+
size 4968488768
|
model-00033-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:71d86d4a8ee5362c6c0691298928002d07cf5e8c2135dcb9ee3c96cd5d7971c5
|
3 |
+
size 4968488768
|
model-00034-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:198c562abe029ace349716692fe4bdc8634eab16aa634a8d47278b90849b3f1c
|
3 |
+
size 4968488768
|
model-00035-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a95189b94f1635bc95ecc83fc8b2c1728a6375e29fb9e403c3f92d43c1858ec4
|
3 |
+
size 4968488832
|
model-00036-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:405db95953601cf83f81f6078f002d01fd122dce624ffc3a52e0f709f589683a
|
3 |
+
size 4968488840
|
model-00037-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fb47d6e1ec60ca9676f957fa407e892da2b238576e764f1f986862d4b1eb0f74
|
3 |
+
size 4968488840
|
model-00038-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:987a027eeaa4014390116458bdbf647b9a6d998b84ceabf685aec9e27aec9eb5
|
3 |
+
size 4968488840
|
model-00039-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:77770955b8fabf33dd2cd7c9890b8dcfa616761da49ce25e06390d3b373e293e
|
3 |
+
size 4968488840
|
model-00040-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3b30b2d7fe43741956ab5656eff4110d4bc72e9c22658f2fca80d30b6c8e1e07
|
3 |
+
size 4968488840
|
model-00041-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3f774b439898bc7682016facabb5a627aaca4871320eecfae31a6c50aef7df33
|
3 |
+
size 4968488840
|
model-00042-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bac5034cb04396084e661e364b26dccb7ee1169d6e2f51aa3ec06940289eaf01
|
3 |
+
size 4968488840
|
model-00043-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9db28d7373a4f6c7f19ef578617637f2e02ae20a3ef365db55db4dae6470e13e
|
3 |
+
size 4218919168
|
model-00044-of-00044.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:56b10aec9a93280980a830629be8161bacec7e0f741f7b87e5fcb3de68104edb
|
3 |
+
size 4202692736
|
model.safetensors.index.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|