peter szemraj commited on
Commit ·
db742b8
1
Parent(s): dfb9989
add sharded checkpoint
Browse files- .gitattributes +7 -0
- config.json +34 -0
- generation_config.json +7 -0
- pip_pkgs.txt +427 -0
- pytorch_model-00001-of-00007.bin +3 -0
- pytorch_model-00002-of-00007.bin +3 -0
- pytorch_model-00003-of-00007.bin +3 -0
- pytorch_model-00004-of-00007.bin +3 -0
- pytorch_model-00005-of-00007.bin +3 -0
- pytorch_model-00006-of-00007.bin +3 -0
- pytorch_model-00007-of-00007.bin +3 -0
- pytorch_model.bin.index.json +690 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,10 @@ 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 |
+
pytorch_model-00003-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
pytorch_model-00004-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
pytorch_model-00005-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
pytorch_model-00006-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
pytorch_model-00007-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
pytorch_model-00001-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
pytorch_model-00002-of-00007.bin filter=lfs diff=lfs merge=lfs -text
|
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openlm-research/open_llama_13b",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"eos_token_id": 2,
|
| 8 |
+
"hidden_act": "silu",
|
| 9 |
+
"hidden_size": 5120,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 13824,
|
| 12 |
+
"max_position_embeddings": 2048,
|
| 13 |
+
"model_type": "llama",
|
| 14 |
+
"num_attention_heads": 40,
|
| 15 |
+
"num_hidden_layers": 40,
|
| 16 |
+
"pad_token_id": 0,
|
| 17 |
+
"quantization_config": {
|
| 18 |
+
"bnb_4bit_compute_dtype": "float32",
|
| 19 |
+
"bnb_4bit_quant_type": "fp4",
|
| 20 |
+
"bnb_4bit_use_double_quant": false,
|
| 21 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
| 22 |
+
"llm_int8_has_fp16_weight": false,
|
| 23 |
+
"llm_int8_skip_modules": null,
|
| 24 |
+
"llm_int8_threshold": 6.0,
|
| 25 |
+
"load_in_4bit": false,
|
| 26 |
+
"load_in_8bit": true
|
| 27 |
+
},
|
| 28 |
+
"rms_norm_eps": 1e-06,
|
| 29 |
+
"tie_word_embeddings": false,
|
| 30 |
+
"torch_dtype": "float16",
|
| 31 |
+
"transformers_version": "4.30.2",
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 32000
|
| 34 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.30.2"
|
| 7 |
+
}
|
pip_pkgs.txt
ADDED
|
@@ -0,0 +1,427 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
absl-py==1.4.0
|
| 2 |
+
accelerate==0.20.3
|
| 3 |
+
aiohttp==3.8.4
|
| 4 |
+
aiosignal==1.3.1
|
| 5 |
+
alabaster==0.7.13
|
| 6 |
+
albumentations==1.2.1
|
| 7 |
+
altair==4.2.2
|
| 8 |
+
anyio==3.7.0
|
| 9 |
+
appdirs==1.4.4
|
| 10 |
+
argon2-cffi==21.3.0
|
| 11 |
+
argon2-cffi-bindings==21.2.0
|
| 12 |
+
array-record==0.4.0
|
| 13 |
+
arviz==0.15.1
|
| 14 |
+
astropy==5.2.2
|
| 15 |
+
astunparse==1.6.3
|
| 16 |
+
async-timeout==4.0.2
|
| 17 |
+
attrs==23.1.0
|
| 18 |
+
audioread==3.0.0
|
| 19 |
+
autograd==1.6.1
|
| 20 |
+
Babel==2.12.1
|
| 21 |
+
backcall==0.2.0
|
| 22 |
+
beautifulsoup4==4.11.2
|
| 23 |
+
bitsandbytes==0.38.0.post2
|
| 24 |
+
bleach==6.0.0
|
| 25 |
+
blis==0.7.9
|
| 26 |
+
blosc2==2.0.0
|
| 27 |
+
bokeh==2.4.3
|
| 28 |
+
branca==0.6.0
|
| 29 |
+
build==0.10.0
|
| 30 |
+
CacheControl==0.13.1
|
| 31 |
+
cached-property==1.5.2
|
| 32 |
+
cachetools==5.3.1
|
| 33 |
+
catalogue==2.0.8
|
| 34 |
+
certifi==2023.5.7
|
| 35 |
+
cffi==1.15.1
|
| 36 |
+
chardet==4.0.0
|
| 37 |
+
charset-normalizer==2.0.12
|
| 38 |
+
chex==0.1.7
|
| 39 |
+
click==8.1.3
|
| 40 |
+
click-plugins==1.1.1
|
| 41 |
+
cligj==0.7.2
|
| 42 |
+
cloudpickle==2.2.1
|
| 43 |
+
cmake==3.25.2
|
| 44 |
+
cmdstanpy==1.1.0
|
| 45 |
+
colorcet==3.0.1
|
| 46 |
+
colorlover==0.3.0
|
| 47 |
+
community==1.0.0b1
|
| 48 |
+
confection==0.0.4
|
| 49 |
+
cons==0.4.6
|
| 50 |
+
contextlib2==0.6.0.post1
|
| 51 |
+
contourpy==1.1.0
|
| 52 |
+
convertdate==2.4.0
|
| 53 |
+
cufflinks==0.17.3
|
| 54 |
+
cupy-cuda11x==11.0.0
|
| 55 |
+
cvxopt==1.3.1
|
| 56 |
+
cvxpy==1.3.1
|
| 57 |
+
cycler==0.11.0
|
| 58 |
+
cymem==2.0.7
|
| 59 |
+
Cython==0.29.35
|
| 60 |
+
dask==2022.12.1
|
| 61 |
+
datascience==0.17.6
|
| 62 |
+
db-dtypes==1.1.1
|
| 63 |
+
dbus-python==1.2.16
|
| 64 |
+
debugpy==1.6.6
|
| 65 |
+
decorator==4.4.2
|
| 66 |
+
defusedxml==0.7.1
|
| 67 |
+
distributed==2022.12.1
|
| 68 |
+
dlib==19.24.2
|
| 69 |
+
dm-tree==0.1.8
|
| 70 |
+
docutils==0.16
|
| 71 |
+
dopamine-rl==4.0.6
|
| 72 |
+
duckdb==0.8.1
|
| 73 |
+
earthengine-api==0.1.357
|
| 74 |
+
easydict==1.10
|
| 75 |
+
ecos==2.0.12
|
| 76 |
+
editdistance==0.6.2
|
| 77 |
+
en-core-web-sm @ https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.5.0/en_core_web_sm-3.5.0-py3-none-any.whl#sha256=0964370218b7e1672a30ac50d72cdc6b16f7c867496f1d60925691188f4d2510
|
| 78 |
+
entrypoints==0.4
|
| 79 |
+
ephem==4.1.4
|
| 80 |
+
et-xmlfile==1.1.0
|
| 81 |
+
etils==1.3.0
|
| 82 |
+
etuples==0.3.9
|
| 83 |
+
exceptiongroup==1.1.1
|
| 84 |
+
fastai==2.7.12
|
| 85 |
+
fastcore==1.5.29
|
| 86 |
+
fastdownload==0.0.7
|
| 87 |
+
fastjsonschema==2.17.1
|
| 88 |
+
fastprogress==1.0.3
|
| 89 |
+
fastrlock==0.8.1
|
| 90 |
+
filelock==3.12.2
|
| 91 |
+
Fiona==1.9.4.post1
|
| 92 |
+
firebase-admin==5.3.0
|
| 93 |
+
Flask==2.2.5
|
| 94 |
+
flatbuffers==23.5.26
|
| 95 |
+
flax==0.6.11
|
| 96 |
+
folium==0.14.0
|
| 97 |
+
fonttools==4.40.0
|
| 98 |
+
frozendict==2.3.8
|
| 99 |
+
frozenlist==1.3.3
|
| 100 |
+
fsspec==2023.6.0
|
| 101 |
+
future==0.18.3
|
| 102 |
+
gast==0.4.0
|
| 103 |
+
gcsfs==2023.6.0
|
| 104 |
+
GDAL==3.3.2
|
| 105 |
+
gdown==4.6.6
|
| 106 |
+
gensim==4.3.1
|
| 107 |
+
geographiclib==2.0
|
| 108 |
+
geopandas==0.13.2
|
| 109 |
+
geopy==2.3.0
|
| 110 |
+
gin-config==0.5.0
|
| 111 |
+
glob2==0.7
|
| 112 |
+
google==2.0.3
|
| 113 |
+
google-api-core==2.11.1
|
| 114 |
+
google-api-python-client==2.84.0
|
| 115 |
+
google-auth==2.17.3
|
| 116 |
+
google-auth-httplib2==0.1.0
|
| 117 |
+
google-auth-oauthlib==1.0.0
|
| 118 |
+
google-cloud-bigquery==3.10.0
|
| 119 |
+
google-cloud-bigquery-connection==1.12.0
|
| 120 |
+
google-cloud-bigquery-storage==2.20.0
|
| 121 |
+
google-cloud-core==2.3.2
|
| 122 |
+
google-cloud-datastore==2.15.2
|
| 123 |
+
google-cloud-firestore==2.11.1
|
| 124 |
+
google-cloud-functions==1.13.0
|
| 125 |
+
google-cloud-language==2.9.1
|
| 126 |
+
google-cloud-storage==2.8.0
|
| 127 |
+
google-cloud-translate==3.11.1
|
| 128 |
+
google-colab @ file:///colabtools/dist/google-colab-1.0.0.tar.gz#sha256=e1dbf1675b82118c05ec840f6e658f6a5329ad5203c0f6051f010bb87ae5147f
|
| 129 |
+
google-crc32c==1.5.0
|
| 130 |
+
google-pasta==0.2.0
|
| 131 |
+
google-resumable-media==2.5.0
|
| 132 |
+
googleapis-common-protos==1.59.1
|
| 133 |
+
googledrivedownloader==0.4
|
| 134 |
+
graphviz==0.20.1
|
| 135 |
+
greenlet==2.0.2
|
| 136 |
+
grpc-google-iam-v1==0.12.6
|
| 137 |
+
grpcio==1.56.0
|
| 138 |
+
grpcio-status==1.48.2
|
| 139 |
+
gspread==3.4.2
|
| 140 |
+
gspread-dataframe==3.0.8
|
| 141 |
+
gym==0.25.2
|
| 142 |
+
gym-notices==0.0.8
|
| 143 |
+
h5netcdf==1.2.0
|
| 144 |
+
h5py==3.8.0
|
| 145 |
+
holidays==0.27.1
|
| 146 |
+
holoviews==1.15.4
|
| 147 |
+
html5lib==1.1
|
| 148 |
+
httpimport==1.3.0
|
| 149 |
+
httplib2==0.21.0
|
| 150 |
+
huggingface-hub==0.15.1
|
| 151 |
+
humanize==4.6.0
|
| 152 |
+
hyperopt==0.2.7
|
| 153 |
+
idna==3.4
|
| 154 |
+
imageio==2.25.1
|
| 155 |
+
imageio-ffmpeg==0.4.8
|
| 156 |
+
imagesize==1.4.1
|
| 157 |
+
imbalanced-learn==0.10.1
|
| 158 |
+
imgaug==0.4.0
|
| 159 |
+
importlib-resources==5.12.0
|
| 160 |
+
imutils==0.5.4
|
| 161 |
+
inflect==6.0.4
|
| 162 |
+
iniconfig==2.0.0
|
| 163 |
+
intel-openmp==2023.1.0
|
| 164 |
+
ipykernel==5.5.6
|
| 165 |
+
ipython==7.34.0
|
| 166 |
+
ipython-genutils==0.2.0
|
| 167 |
+
ipython-sql==0.4.1
|
| 168 |
+
ipywidgets==7.7.1
|
| 169 |
+
itsdangerous==2.1.2
|
| 170 |
+
jax==0.4.10
|
| 171 |
+
jaxlib @ https://storage.googleapis.com/jax-releases/cuda11/jaxlib-0.4.10+cuda11.cudnn86-cp310-cp310-manylinux2014_x86_64.whl#sha256=fe53205ef12727c80ed5ac2d4506d6732c0c3db69ede4565a7d4df98e609af84
|
| 172 |
+
jieba==0.42.1
|
| 173 |
+
Jinja2==3.1.2
|
| 174 |
+
joblib==1.2.0
|
| 175 |
+
jsonpickle==3.0.1
|
| 176 |
+
jsonschema==4.3.3
|
| 177 |
+
jupyter-client==6.1.12
|
| 178 |
+
jupyter-console==6.1.0
|
| 179 |
+
jupyter-server==1.24.0
|
| 180 |
+
jupyter_core==5.3.1
|
| 181 |
+
jupyterlab-pygments==0.2.2
|
| 182 |
+
jupyterlab-widgets==3.0.7
|
| 183 |
+
kaggle==1.5.13
|
| 184 |
+
keras==2.12.0
|
| 185 |
+
kiwisolver==1.4.4
|
| 186 |
+
langcodes==3.3.0
|
| 187 |
+
lazy_loader==0.2
|
| 188 |
+
libclang==16.0.0
|
| 189 |
+
librosa==0.10.0.post2
|
| 190 |
+
lightgbm==3.3.5
|
| 191 |
+
lit==16.0.6
|
| 192 |
+
llvmlite==0.39.1
|
| 193 |
+
locket==1.0.0
|
| 194 |
+
logical-unification==0.4.6
|
| 195 |
+
LunarCalendar==0.0.9
|
| 196 |
+
lxml==4.9.2
|
| 197 |
+
Markdown==3.4.3
|
| 198 |
+
markdown-it-py==3.0.0
|
| 199 |
+
MarkupSafe==2.1.3
|
| 200 |
+
matplotlib==3.7.1
|
| 201 |
+
matplotlib-inline==0.1.6
|
| 202 |
+
matplotlib-venn==0.11.9
|
| 203 |
+
mdurl==0.1.2
|
| 204 |
+
miniKanren==1.0.3
|
| 205 |
+
missingno==0.5.2
|
| 206 |
+
mistune==0.8.4
|
| 207 |
+
mizani==0.8.1
|
| 208 |
+
mkl==2019.0
|
| 209 |
+
ml-dtypes==0.2.0
|
| 210 |
+
mlxtend==0.14.0
|
| 211 |
+
more-itertools==9.1.0
|
| 212 |
+
moviepy==1.0.3
|
| 213 |
+
mpmath==1.3.0
|
| 214 |
+
msgpack==1.0.5
|
| 215 |
+
multidict==6.0.4
|
| 216 |
+
multipledispatch==0.6.0
|
| 217 |
+
multitasking==0.0.11
|
| 218 |
+
murmurhash==1.0.9
|
| 219 |
+
music21==8.1.0
|
| 220 |
+
natsort==8.3.1
|
| 221 |
+
nbclient==0.8.0
|
| 222 |
+
nbconvert==6.5.4
|
| 223 |
+
nbformat==5.9.0
|
| 224 |
+
nest-asyncio==1.5.6
|
| 225 |
+
networkx==3.1
|
| 226 |
+
nibabel==3.0.2
|
| 227 |
+
nltk==3.8.1
|
| 228 |
+
notebook==6.4.8
|
| 229 |
+
numba==0.56.4
|
| 230 |
+
numexpr==2.8.4
|
| 231 |
+
numpy==1.22.4
|
| 232 |
+
oauth2client==4.1.3
|
| 233 |
+
oauthlib==3.2.2
|
| 234 |
+
opencv-contrib-python==4.7.0.72
|
| 235 |
+
opencv-python==4.7.0.72
|
| 236 |
+
opencv-python-headless==4.7.0.72
|
| 237 |
+
openpyxl==3.0.10
|
| 238 |
+
opt-einsum==3.3.0
|
| 239 |
+
optax==0.1.5
|
| 240 |
+
orbax-checkpoint==0.2.6
|
| 241 |
+
osqp==0.6.2.post8
|
| 242 |
+
packaging==23.1
|
| 243 |
+
palettable==3.3.3
|
| 244 |
+
pandas==1.5.3
|
| 245 |
+
pandas-datareader==0.10.0
|
| 246 |
+
pandas-gbq==0.17.9
|
| 247 |
+
pandocfilters==1.5.0
|
| 248 |
+
panel==0.14.4
|
| 249 |
+
param==1.13.0
|
| 250 |
+
parso==0.8.3
|
| 251 |
+
partd==1.4.0
|
| 252 |
+
pathlib==1.0.1
|
| 253 |
+
pathy==0.10.2
|
| 254 |
+
patsy==0.5.3
|
| 255 |
+
pexpect==4.8.0
|
| 256 |
+
pickleshare==0.7.5
|
| 257 |
+
Pillow==8.4.0
|
| 258 |
+
pip-tools==6.13.0
|
| 259 |
+
platformdirs==3.7.0
|
| 260 |
+
plotly==5.13.1
|
| 261 |
+
plotnine==0.10.1
|
| 262 |
+
pluggy==1.2.0
|
| 263 |
+
polars==0.17.3
|
| 264 |
+
pooch==1.6.0
|
| 265 |
+
portpicker==1.5.2
|
| 266 |
+
prefetch-generator==1.0.3
|
| 267 |
+
preshed==3.0.8
|
| 268 |
+
prettytable==0.7.2
|
| 269 |
+
proglog==0.1.10
|
| 270 |
+
progressbar2==4.2.0
|
| 271 |
+
prometheus-client==0.17.0
|
| 272 |
+
promise==2.3
|
| 273 |
+
prompt-toolkit==3.0.38
|
| 274 |
+
prophet==1.1.4
|
| 275 |
+
proto-plus==1.22.3
|
| 276 |
+
protobuf==3.20.3
|
| 277 |
+
psutil==5.9.5
|
| 278 |
+
psycopg2==2.9.6
|
| 279 |
+
ptyprocess==0.7.0
|
| 280 |
+
py-cpuinfo==9.0.0
|
| 281 |
+
py4j==0.10.9.7
|
| 282 |
+
pyarrow==9.0.0
|
| 283 |
+
pyasn1==0.5.0
|
| 284 |
+
pyasn1-modules==0.3.0
|
| 285 |
+
pycocotools==2.0.6
|
| 286 |
+
pycparser==2.21
|
| 287 |
+
pyct==0.5.0
|
| 288 |
+
pydantic==1.10.9
|
| 289 |
+
pydata-google-auth==1.8.0
|
| 290 |
+
pydot==1.4.2
|
| 291 |
+
pydot-ng==2.0.0
|
| 292 |
+
pydotplus==2.0.2
|
| 293 |
+
PyDrive==1.3.1
|
| 294 |
+
pyerfa==2.0.0.3
|
| 295 |
+
pygame==2.4.0
|
| 296 |
+
Pygments==2.14.0
|
| 297 |
+
PyGObject==3.36.0
|
| 298 |
+
pymc==5.1.2
|
| 299 |
+
PyMeeus==0.5.12
|
| 300 |
+
pymystem3==0.2.0
|
| 301 |
+
PyOpenGL==3.1.7
|
| 302 |
+
pyparsing==3.1.0
|
| 303 |
+
pyproj==3.6.0
|
| 304 |
+
pyproject_hooks==1.0.0
|
| 305 |
+
pyrsistent==0.19.3
|
| 306 |
+
PySocks==1.7.1
|
| 307 |
+
pytensor==2.10.1
|
| 308 |
+
pytest==7.2.2
|
| 309 |
+
python-apt==0.0.0
|
| 310 |
+
python-dateutil==2.8.2
|
| 311 |
+
python-louvain==0.16
|
| 312 |
+
python-slugify==8.0.1
|
| 313 |
+
python-utils==3.7.0
|
| 314 |
+
pytz==2022.7.1
|
| 315 |
+
pyviz-comms==2.3.2
|
| 316 |
+
PyWavelets==1.4.1
|
| 317 |
+
PyYAML==6.0
|
| 318 |
+
pyzmq==23.2.1
|
| 319 |
+
qdldl==0.1.7
|
| 320 |
+
qudida==0.0.4
|
| 321 |
+
regex==2022.10.31
|
| 322 |
+
requests==2.27.1
|
| 323 |
+
requests-oauthlib==1.3.1
|
| 324 |
+
requests-unixsocket==0.2.0
|
| 325 |
+
requirements-parser==0.5.0
|
| 326 |
+
rich==13.4.2
|
| 327 |
+
rpy2==3.5.5
|
| 328 |
+
rsa==4.9
|
| 329 |
+
safetensors==0.3.1
|
| 330 |
+
scikit-image==0.19.3
|
| 331 |
+
scikit-learn==1.2.2
|
| 332 |
+
scipy==1.10.1
|
| 333 |
+
scs==3.2.3
|
| 334 |
+
seaborn==0.12.2
|
| 335 |
+
Send2Trash==1.8.2
|
| 336 |
+
sentencepiece==0.1.99
|
| 337 |
+
shapely==2.0.1
|
| 338 |
+
six==1.16.0
|
| 339 |
+
sklearn-pandas==2.2.0
|
| 340 |
+
smart-open==6.3.0
|
| 341 |
+
sniffio==1.3.0
|
| 342 |
+
snowballstemmer==2.2.0
|
| 343 |
+
sortedcontainers==2.4.0
|
| 344 |
+
soundfile==0.12.1
|
| 345 |
+
soupsieve==2.4.1
|
| 346 |
+
soxr==0.3.5
|
| 347 |
+
spacy==3.5.3
|
| 348 |
+
spacy-legacy==3.0.12
|
| 349 |
+
spacy-loggers==1.0.4
|
| 350 |
+
Sphinx==3.5.4
|
| 351 |
+
sphinxcontrib-applehelp==1.0.4
|
| 352 |
+
sphinxcontrib-devhelp==1.0.2
|
| 353 |
+
sphinxcontrib-htmlhelp==2.0.1
|
| 354 |
+
sphinxcontrib-jsmath==1.0.1
|
| 355 |
+
sphinxcontrib-qthelp==1.0.3
|
| 356 |
+
sphinxcontrib-serializinghtml==1.1.5
|
| 357 |
+
SQLAlchemy==2.0.16
|
| 358 |
+
sqlparse==0.4.4
|
| 359 |
+
srsly==2.4.6
|
| 360 |
+
statsmodels==0.13.5
|
| 361 |
+
sympy==1.11.1
|
| 362 |
+
tables==3.8.0
|
| 363 |
+
tabulate==0.8.10
|
| 364 |
+
tblib==2.0.0
|
| 365 |
+
tenacity==8.2.2
|
| 366 |
+
tensorboard==2.12.3
|
| 367 |
+
tensorboard-data-server==0.7.1
|
| 368 |
+
tensorflow==2.12.0
|
| 369 |
+
tensorflow-datasets==4.9.2
|
| 370 |
+
tensorflow-estimator==2.12.0
|
| 371 |
+
tensorflow-gcs-config==2.12.0
|
| 372 |
+
tensorflow-hub==0.13.0
|
| 373 |
+
tensorflow-io-gcs-filesystem==0.32.0
|
| 374 |
+
tensorflow-metadata==1.13.1
|
| 375 |
+
tensorflow-probability==0.20.1
|
| 376 |
+
tensorstore==0.1.38
|
| 377 |
+
termcolor==2.3.0
|
| 378 |
+
terminado==0.17.1
|
| 379 |
+
text-unidecode==1.3
|
| 380 |
+
textblob==0.17.1
|
| 381 |
+
tf-slim==1.1.0
|
| 382 |
+
thinc==8.1.10
|
| 383 |
+
threadpoolctl==3.1.0
|
| 384 |
+
tifffile==2023.4.12
|
| 385 |
+
tiktoken==0.4.0
|
| 386 |
+
tinycss2==1.2.1
|
| 387 |
+
tokenizers==0.13.3
|
| 388 |
+
toml==0.10.2
|
| 389 |
+
tomli==2.0.1
|
| 390 |
+
toolz==0.12.0
|
| 391 |
+
torch @ https://download.pytorch.org/whl/cu118/torch-2.0.1%2Bcu118-cp310-cp310-linux_x86_64.whl#sha256=a7a49d459bf4862f64f7bc1a68beccf8881c2fa9f3e0569608e16ba6f85ebf7b
|
| 392 |
+
torchaudio @ https://download.pytorch.org/whl/cu118/torchaudio-2.0.2%2Bcu118-cp310-cp310-linux_x86_64.whl#sha256=26692645ea061a005c57ec581a2d0425210ac6ba9f923edf11cc9b0ef3a111e9
|
| 393 |
+
torchdata==0.6.1
|
| 394 |
+
torchsummary==1.5.1
|
| 395 |
+
torchtext==0.15.2
|
| 396 |
+
torchvision @ https://download.pytorch.org/whl/cu118/torchvision-0.15.2%2Bcu118-cp310-cp310-linux_x86_64.whl#sha256=19ca4ab5d6179bbe53cff79df1a855ee6533c2861ddc7389f68349d8b9f8302a
|
| 397 |
+
tornado==6.3.1
|
| 398 |
+
tqdm==4.65.0
|
| 399 |
+
traitlets==5.7.1
|
| 400 |
+
transformers==4.30.2
|
| 401 |
+
triton==2.0.0
|
| 402 |
+
tweepy==4.13.0
|
| 403 |
+
typer==0.7.0
|
| 404 |
+
types-setuptools==68.0.0.0
|
| 405 |
+
typing_extensions==4.6.3
|
| 406 |
+
tzlocal==5.0.1
|
| 407 |
+
uritemplate==4.1.1
|
| 408 |
+
urllib3==1.26.16
|
| 409 |
+
vega-datasets==0.9.0
|
| 410 |
+
wasabi==1.1.2
|
| 411 |
+
wcwidth==0.2.6
|
| 412 |
+
webcolors==1.13
|
| 413 |
+
webencodings==0.5.1
|
| 414 |
+
websocket-client==1.6.0
|
| 415 |
+
Werkzeug==2.3.6
|
| 416 |
+
widgetsnbextension==3.6.4
|
| 417 |
+
wordcloud==1.8.2.2
|
| 418 |
+
wrapt==1.14.1
|
| 419 |
+
xarray==2022.12.0
|
| 420 |
+
xarray-einstats==0.5.1
|
| 421 |
+
xgboost==1.7.6
|
| 422 |
+
xlrd==2.0.1
|
| 423 |
+
yarl==1.9.2
|
| 424 |
+
yellowbrick==1.5
|
| 425 |
+
yfinance==0.2.21
|
| 426 |
+
zict==3.0.0
|
| 427 |
+
zipp==3.15.0
|
pytorch_model-00001-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:671088665cfe94a581a95bb6be13c58a23a5ae5a027438bc81031b235856927d
|
| 3 |
+
size 1993555591
|
pytorch_model-00002-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:efb1b24285eac61dc21841d9ff843e7e32e19806d2e79375f1ea12b7ee7f2aa9
|
| 3 |
+
size 1930838326
|
pytorch_model-00003-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12cbf59a383fb62409eb28fec3d4a53d6c4b871abc73d82dcc9d5020813674eb
|
| 3 |
+
size 1975437219
|
pytorch_model-00004-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d9ad9b0cc0470a802337649a21d192b76d76dd115475963d66a79e93dd3909e
|
| 3 |
+
size 1975402403
|
pytorch_model-00005-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c45b91e21569477bd8d67422a33445bc102279de75283ceb2204440b5d1e4af7
|
| 3 |
+
size 1975458441
|
pytorch_model-00006-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5eb018aee85017e746c052ae0a8e779fb26e81cc445241df387d10ce17d57b32
|
| 3 |
+
size 1983309103
|
pytorch_model-00007-of-00007.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0004b7bab8bb63915cf230092a88ef35c2c5863b402ed80e0f45f3f536e0c894
|
| 3 |
+
size 1518719055
|
pytorch_model.bin.index.json
ADDED
|
@@ -0,0 +1,690 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_size": 13352483840
|
| 4 |
+
},
|
| 5 |
+
"weight_map": {
|
| 6 |
+
"lm_head.weight": "pytorch_model-00007-of-00007.bin",
|
| 7 |
+
"model.embed_tokens.weight": "pytorch_model-00001-of-00007.bin",
|
| 8 |
+
"model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 9 |
+
"model.layers.0.mlp.down_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 10 |
+
"model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 11 |
+
"model.layers.0.mlp.gate_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 12 |
+
"model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 13 |
+
"model.layers.0.mlp.up_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 14 |
+
"model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 15 |
+
"model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 16 |
+
"model.layers.0.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 17 |
+
"model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 18 |
+
"model.layers.0.self_attn.o_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 19 |
+
"model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 20 |
+
"model.layers.0.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 21 |
+
"model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 22 |
+
"model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00007.bin",
|
| 23 |
+
"model.layers.0.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 24 |
+
"model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 25 |
+
"model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 26 |
+
"model.layers.1.mlp.down_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 27 |
+
"model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 28 |
+
"model.layers.1.mlp.gate_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 29 |
+
"model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 30 |
+
"model.layers.1.mlp.up_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 31 |
+
"model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 32 |
+
"model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 33 |
+
"model.layers.1.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 34 |
+
"model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 35 |
+
"model.layers.1.self_attn.o_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 36 |
+
"model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 37 |
+
"model.layers.1.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 38 |
+
"model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 39 |
+
"model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00007.bin",
|
| 40 |
+
"model.layers.1.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 41 |
+
"model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 42 |
+
"model.layers.10.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 43 |
+
"model.layers.10.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 44 |
+
"model.layers.10.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 45 |
+
"model.layers.10.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 46 |
+
"model.layers.10.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 47 |
+
"model.layers.10.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 48 |
+
"model.layers.10.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 49 |
+
"model.layers.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 50 |
+
"model.layers.10.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 51 |
+
"model.layers.10.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 52 |
+
"model.layers.10.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 53 |
+
"model.layers.10.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 54 |
+
"model.layers.10.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 55 |
+
"model.layers.10.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 56 |
+
"model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 57 |
+
"model.layers.10.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 58 |
+
"model.layers.10.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 59 |
+
"model.layers.11.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 60 |
+
"model.layers.11.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 61 |
+
"model.layers.11.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 62 |
+
"model.layers.11.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 63 |
+
"model.layers.11.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 64 |
+
"model.layers.11.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 65 |
+
"model.layers.11.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 66 |
+
"model.layers.11.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 67 |
+
"model.layers.11.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 68 |
+
"model.layers.11.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 69 |
+
"model.layers.11.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 70 |
+
"model.layers.11.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 71 |
+
"model.layers.11.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 72 |
+
"model.layers.11.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 73 |
+
"model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 74 |
+
"model.layers.11.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 75 |
+
"model.layers.11.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 76 |
+
"model.layers.12.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 77 |
+
"model.layers.12.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 78 |
+
"model.layers.12.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 79 |
+
"model.layers.12.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 80 |
+
"model.layers.12.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 81 |
+
"model.layers.12.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 82 |
+
"model.layers.12.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 83 |
+
"model.layers.12.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 84 |
+
"model.layers.12.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 85 |
+
"model.layers.12.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 86 |
+
"model.layers.12.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 87 |
+
"model.layers.12.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 88 |
+
"model.layers.12.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 89 |
+
"model.layers.12.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 90 |
+
"model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 91 |
+
"model.layers.12.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 92 |
+
"model.layers.12.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 93 |
+
"model.layers.13.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 94 |
+
"model.layers.13.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 95 |
+
"model.layers.13.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 96 |
+
"model.layers.13.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 97 |
+
"model.layers.13.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 98 |
+
"model.layers.13.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 99 |
+
"model.layers.13.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 100 |
+
"model.layers.13.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 101 |
+
"model.layers.13.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 102 |
+
"model.layers.13.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 103 |
+
"model.layers.13.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 104 |
+
"model.layers.13.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 105 |
+
"model.layers.13.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 106 |
+
"model.layers.13.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 107 |
+
"model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 108 |
+
"model.layers.13.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 109 |
+
"model.layers.13.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 110 |
+
"model.layers.14.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 111 |
+
"model.layers.14.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 112 |
+
"model.layers.14.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 113 |
+
"model.layers.14.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 114 |
+
"model.layers.14.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 115 |
+
"model.layers.14.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 116 |
+
"model.layers.14.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 117 |
+
"model.layers.14.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 118 |
+
"model.layers.14.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 119 |
+
"model.layers.14.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 120 |
+
"model.layers.14.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 121 |
+
"model.layers.14.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 122 |
+
"model.layers.14.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 123 |
+
"model.layers.14.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 124 |
+
"model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 125 |
+
"model.layers.14.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 126 |
+
"model.layers.14.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 127 |
+
"model.layers.15.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 128 |
+
"model.layers.15.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 129 |
+
"model.layers.15.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 130 |
+
"model.layers.15.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 131 |
+
"model.layers.15.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 132 |
+
"model.layers.15.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 133 |
+
"model.layers.15.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 134 |
+
"model.layers.15.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 135 |
+
"model.layers.15.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 136 |
+
"model.layers.15.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 137 |
+
"model.layers.15.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 138 |
+
"model.layers.15.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 139 |
+
"model.layers.15.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 140 |
+
"model.layers.15.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 141 |
+
"model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 142 |
+
"model.layers.15.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 143 |
+
"model.layers.15.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 144 |
+
"model.layers.16.input_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 145 |
+
"model.layers.16.mlp.down_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 146 |
+
"model.layers.16.mlp.down_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 147 |
+
"model.layers.16.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 148 |
+
"model.layers.16.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 149 |
+
"model.layers.16.mlp.up_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 150 |
+
"model.layers.16.mlp.up_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 151 |
+
"model.layers.16.post_attention_layernorm.weight": "pytorch_model-00003-of-00007.bin",
|
| 152 |
+
"model.layers.16.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 153 |
+
"model.layers.16.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 154 |
+
"model.layers.16.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 155 |
+
"model.layers.16.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 156 |
+
"model.layers.16.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 157 |
+
"model.layers.16.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 158 |
+
"model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 159 |
+
"model.layers.16.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 160 |
+
"model.layers.16.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 161 |
+
"model.layers.17.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 162 |
+
"model.layers.17.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 163 |
+
"model.layers.17.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 164 |
+
"model.layers.17.mlp.gate_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 165 |
+
"model.layers.17.mlp.gate_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 166 |
+
"model.layers.17.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 167 |
+
"model.layers.17.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 168 |
+
"model.layers.17.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 169 |
+
"model.layers.17.self_attn.k_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 170 |
+
"model.layers.17.self_attn.k_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 171 |
+
"model.layers.17.self_attn.o_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 172 |
+
"model.layers.17.self_attn.o_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 173 |
+
"model.layers.17.self_attn.q_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 174 |
+
"model.layers.17.self_attn.q_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 175 |
+
"model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00007.bin",
|
| 176 |
+
"model.layers.17.self_attn.v_proj.SCB": "pytorch_model-00003-of-00007.bin",
|
| 177 |
+
"model.layers.17.self_attn.v_proj.weight": "pytorch_model-00003-of-00007.bin",
|
| 178 |
+
"model.layers.18.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 179 |
+
"model.layers.18.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 180 |
+
"model.layers.18.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 181 |
+
"model.layers.18.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 182 |
+
"model.layers.18.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 183 |
+
"model.layers.18.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 184 |
+
"model.layers.18.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 185 |
+
"model.layers.18.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 186 |
+
"model.layers.18.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 187 |
+
"model.layers.18.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 188 |
+
"model.layers.18.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 189 |
+
"model.layers.18.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 190 |
+
"model.layers.18.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 191 |
+
"model.layers.18.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 192 |
+
"model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 193 |
+
"model.layers.18.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 194 |
+
"model.layers.18.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 195 |
+
"model.layers.19.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 196 |
+
"model.layers.19.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 197 |
+
"model.layers.19.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 198 |
+
"model.layers.19.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 199 |
+
"model.layers.19.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 200 |
+
"model.layers.19.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 201 |
+
"model.layers.19.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 202 |
+
"model.layers.19.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 203 |
+
"model.layers.19.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 204 |
+
"model.layers.19.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 205 |
+
"model.layers.19.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 206 |
+
"model.layers.19.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 207 |
+
"model.layers.19.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 208 |
+
"model.layers.19.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 209 |
+
"model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 210 |
+
"model.layers.19.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 211 |
+
"model.layers.19.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 212 |
+
"model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 213 |
+
"model.layers.2.mlp.down_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 214 |
+
"model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 215 |
+
"model.layers.2.mlp.gate_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 216 |
+
"model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 217 |
+
"model.layers.2.mlp.up_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 218 |
+
"model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 219 |
+
"model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 220 |
+
"model.layers.2.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 221 |
+
"model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 222 |
+
"model.layers.2.self_attn.o_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 223 |
+
"model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 224 |
+
"model.layers.2.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 225 |
+
"model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 226 |
+
"model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00007.bin",
|
| 227 |
+
"model.layers.2.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 228 |
+
"model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 229 |
+
"model.layers.20.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 230 |
+
"model.layers.20.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 231 |
+
"model.layers.20.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 232 |
+
"model.layers.20.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 233 |
+
"model.layers.20.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 234 |
+
"model.layers.20.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 235 |
+
"model.layers.20.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 236 |
+
"model.layers.20.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 237 |
+
"model.layers.20.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 238 |
+
"model.layers.20.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 239 |
+
"model.layers.20.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 240 |
+
"model.layers.20.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 241 |
+
"model.layers.20.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 242 |
+
"model.layers.20.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 243 |
+
"model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 244 |
+
"model.layers.20.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 245 |
+
"model.layers.20.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 246 |
+
"model.layers.21.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 247 |
+
"model.layers.21.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 248 |
+
"model.layers.21.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 249 |
+
"model.layers.21.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 250 |
+
"model.layers.21.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 251 |
+
"model.layers.21.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 252 |
+
"model.layers.21.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 253 |
+
"model.layers.21.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 254 |
+
"model.layers.21.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 255 |
+
"model.layers.21.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 256 |
+
"model.layers.21.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 257 |
+
"model.layers.21.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 258 |
+
"model.layers.21.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 259 |
+
"model.layers.21.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 260 |
+
"model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 261 |
+
"model.layers.21.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 262 |
+
"model.layers.21.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 263 |
+
"model.layers.22.input_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 264 |
+
"model.layers.22.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 265 |
+
"model.layers.22.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 266 |
+
"model.layers.22.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 267 |
+
"model.layers.22.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 268 |
+
"model.layers.22.mlp.up_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 269 |
+
"model.layers.22.mlp.up_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 270 |
+
"model.layers.22.post_attention_layernorm.weight": "pytorch_model-00004-of-00007.bin",
|
| 271 |
+
"model.layers.22.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 272 |
+
"model.layers.22.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 273 |
+
"model.layers.22.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 274 |
+
"model.layers.22.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 275 |
+
"model.layers.22.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 276 |
+
"model.layers.22.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 277 |
+
"model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 278 |
+
"model.layers.22.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 279 |
+
"model.layers.22.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 280 |
+
"model.layers.23.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 281 |
+
"model.layers.23.mlp.down_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 282 |
+
"model.layers.23.mlp.down_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 283 |
+
"model.layers.23.mlp.gate_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 284 |
+
"model.layers.23.mlp.gate_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 285 |
+
"model.layers.23.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 286 |
+
"model.layers.23.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 287 |
+
"model.layers.23.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 288 |
+
"model.layers.23.self_attn.k_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 289 |
+
"model.layers.23.self_attn.k_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 290 |
+
"model.layers.23.self_attn.o_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 291 |
+
"model.layers.23.self_attn.o_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 292 |
+
"model.layers.23.self_attn.q_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 293 |
+
"model.layers.23.self_attn.q_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 294 |
+
"model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00007.bin",
|
| 295 |
+
"model.layers.23.self_attn.v_proj.SCB": "pytorch_model-00004-of-00007.bin",
|
| 296 |
+
"model.layers.23.self_attn.v_proj.weight": "pytorch_model-00004-of-00007.bin",
|
| 297 |
+
"model.layers.24.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 298 |
+
"model.layers.24.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 299 |
+
"model.layers.24.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 300 |
+
"model.layers.24.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 301 |
+
"model.layers.24.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 302 |
+
"model.layers.24.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 303 |
+
"model.layers.24.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 304 |
+
"model.layers.24.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 305 |
+
"model.layers.24.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 306 |
+
"model.layers.24.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 307 |
+
"model.layers.24.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 308 |
+
"model.layers.24.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 309 |
+
"model.layers.24.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 310 |
+
"model.layers.24.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 311 |
+
"model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 312 |
+
"model.layers.24.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 313 |
+
"model.layers.24.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 314 |
+
"model.layers.25.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 315 |
+
"model.layers.25.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 316 |
+
"model.layers.25.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 317 |
+
"model.layers.25.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 318 |
+
"model.layers.25.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 319 |
+
"model.layers.25.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 320 |
+
"model.layers.25.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 321 |
+
"model.layers.25.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 322 |
+
"model.layers.25.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 323 |
+
"model.layers.25.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 324 |
+
"model.layers.25.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 325 |
+
"model.layers.25.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 326 |
+
"model.layers.25.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 327 |
+
"model.layers.25.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 328 |
+
"model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 329 |
+
"model.layers.25.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 330 |
+
"model.layers.25.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 331 |
+
"model.layers.26.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 332 |
+
"model.layers.26.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 333 |
+
"model.layers.26.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 334 |
+
"model.layers.26.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 335 |
+
"model.layers.26.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 336 |
+
"model.layers.26.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 337 |
+
"model.layers.26.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 338 |
+
"model.layers.26.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 339 |
+
"model.layers.26.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 340 |
+
"model.layers.26.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 341 |
+
"model.layers.26.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 342 |
+
"model.layers.26.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 343 |
+
"model.layers.26.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 344 |
+
"model.layers.26.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 345 |
+
"model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 346 |
+
"model.layers.26.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 347 |
+
"model.layers.26.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 348 |
+
"model.layers.27.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 349 |
+
"model.layers.27.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 350 |
+
"model.layers.27.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 351 |
+
"model.layers.27.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 352 |
+
"model.layers.27.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 353 |
+
"model.layers.27.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 354 |
+
"model.layers.27.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 355 |
+
"model.layers.27.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 356 |
+
"model.layers.27.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 357 |
+
"model.layers.27.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 358 |
+
"model.layers.27.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 359 |
+
"model.layers.27.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 360 |
+
"model.layers.27.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 361 |
+
"model.layers.27.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 362 |
+
"model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 363 |
+
"model.layers.27.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 364 |
+
"model.layers.27.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 365 |
+
"model.layers.28.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 366 |
+
"model.layers.28.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 367 |
+
"model.layers.28.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 368 |
+
"model.layers.28.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 369 |
+
"model.layers.28.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 370 |
+
"model.layers.28.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 371 |
+
"model.layers.28.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 372 |
+
"model.layers.28.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 373 |
+
"model.layers.28.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 374 |
+
"model.layers.28.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 375 |
+
"model.layers.28.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 376 |
+
"model.layers.28.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 377 |
+
"model.layers.28.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 378 |
+
"model.layers.28.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 379 |
+
"model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 380 |
+
"model.layers.28.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 381 |
+
"model.layers.28.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 382 |
+
"model.layers.29.input_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 383 |
+
"model.layers.29.mlp.down_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 384 |
+
"model.layers.29.mlp.down_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 385 |
+
"model.layers.29.mlp.gate_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 386 |
+
"model.layers.29.mlp.gate_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 387 |
+
"model.layers.29.mlp.up_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 388 |
+
"model.layers.29.mlp.up_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 389 |
+
"model.layers.29.post_attention_layernorm.weight": "pytorch_model-00005-of-00007.bin",
|
| 390 |
+
"model.layers.29.self_attn.k_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 391 |
+
"model.layers.29.self_attn.k_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 392 |
+
"model.layers.29.self_attn.o_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 393 |
+
"model.layers.29.self_attn.o_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 394 |
+
"model.layers.29.self_attn.q_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 395 |
+
"model.layers.29.self_attn.q_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 396 |
+
"model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00007.bin",
|
| 397 |
+
"model.layers.29.self_attn.v_proj.SCB": "pytorch_model-00005-of-00007.bin",
|
| 398 |
+
"model.layers.29.self_attn.v_proj.weight": "pytorch_model-00005-of-00007.bin",
|
| 399 |
+
"model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 400 |
+
"model.layers.3.mlp.down_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 401 |
+
"model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 402 |
+
"model.layers.3.mlp.gate_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 403 |
+
"model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 404 |
+
"model.layers.3.mlp.up_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 405 |
+
"model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 406 |
+
"model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 407 |
+
"model.layers.3.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 408 |
+
"model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 409 |
+
"model.layers.3.self_attn.o_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 410 |
+
"model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 411 |
+
"model.layers.3.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 412 |
+
"model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 413 |
+
"model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00007.bin",
|
| 414 |
+
"model.layers.3.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 415 |
+
"model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 416 |
+
"model.layers.30.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 417 |
+
"model.layers.30.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 418 |
+
"model.layers.30.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 419 |
+
"model.layers.30.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 420 |
+
"model.layers.30.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 421 |
+
"model.layers.30.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 422 |
+
"model.layers.30.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 423 |
+
"model.layers.30.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 424 |
+
"model.layers.30.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 425 |
+
"model.layers.30.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 426 |
+
"model.layers.30.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 427 |
+
"model.layers.30.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 428 |
+
"model.layers.30.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 429 |
+
"model.layers.30.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 430 |
+
"model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 431 |
+
"model.layers.30.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 432 |
+
"model.layers.30.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 433 |
+
"model.layers.31.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 434 |
+
"model.layers.31.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 435 |
+
"model.layers.31.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 436 |
+
"model.layers.31.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 437 |
+
"model.layers.31.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 438 |
+
"model.layers.31.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 439 |
+
"model.layers.31.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 440 |
+
"model.layers.31.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 441 |
+
"model.layers.31.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 442 |
+
"model.layers.31.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 443 |
+
"model.layers.31.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 444 |
+
"model.layers.31.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 445 |
+
"model.layers.31.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 446 |
+
"model.layers.31.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 447 |
+
"model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 448 |
+
"model.layers.31.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 449 |
+
"model.layers.31.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 450 |
+
"model.layers.32.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 451 |
+
"model.layers.32.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 452 |
+
"model.layers.32.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 453 |
+
"model.layers.32.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 454 |
+
"model.layers.32.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 455 |
+
"model.layers.32.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 456 |
+
"model.layers.32.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 457 |
+
"model.layers.32.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 458 |
+
"model.layers.32.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 459 |
+
"model.layers.32.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 460 |
+
"model.layers.32.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 461 |
+
"model.layers.32.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 462 |
+
"model.layers.32.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 463 |
+
"model.layers.32.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 464 |
+
"model.layers.32.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 465 |
+
"model.layers.32.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 466 |
+
"model.layers.32.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 467 |
+
"model.layers.33.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 468 |
+
"model.layers.33.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 469 |
+
"model.layers.33.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 470 |
+
"model.layers.33.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 471 |
+
"model.layers.33.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 472 |
+
"model.layers.33.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 473 |
+
"model.layers.33.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 474 |
+
"model.layers.33.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 475 |
+
"model.layers.33.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 476 |
+
"model.layers.33.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 477 |
+
"model.layers.33.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 478 |
+
"model.layers.33.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 479 |
+
"model.layers.33.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 480 |
+
"model.layers.33.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 481 |
+
"model.layers.33.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 482 |
+
"model.layers.33.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 483 |
+
"model.layers.33.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 484 |
+
"model.layers.34.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 485 |
+
"model.layers.34.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 486 |
+
"model.layers.34.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 487 |
+
"model.layers.34.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 488 |
+
"model.layers.34.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 489 |
+
"model.layers.34.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 490 |
+
"model.layers.34.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 491 |
+
"model.layers.34.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 492 |
+
"model.layers.34.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 493 |
+
"model.layers.34.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 494 |
+
"model.layers.34.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 495 |
+
"model.layers.34.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 496 |
+
"model.layers.34.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 497 |
+
"model.layers.34.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 498 |
+
"model.layers.34.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 499 |
+
"model.layers.34.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 500 |
+
"model.layers.34.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 501 |
+
"model.layers.35.input_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 502 |
+
"model.layers.35.mlp.down_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 503 |
+
"model.layers.35.mlp.down_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 504 |
+
"model.layers.35.mlp.gate_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 505 |
+
"model.layers.35.mlp.gate_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 506 |
+
"model.layers.35.mlp.up_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 507 |
+
"model.layers.35.mlp.up_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 508 |
+
"model.layers.35.post_attention_layernorm.weight": "pytorch_model-00006-of-00007.bin",
|
| 509 |
+
"model.layers.35.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 510 |
+
"model.layers.35.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 511 |
+
"model.layers.35.self_attn.o_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 512 |
+
"model.layers.35.self_attn.o_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 513 |
+
"model.layers.35.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 514 |
+
"model.layers.35.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 515 |
+
"model.layers.35.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00007.bin",
|
| 516 |
+
"model.layers.35.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 517 |
+
"model.layers.35.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 518 |
+
"model.layers.36.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 519 |
+
"model.layers.36.mlp.down_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 520 |
+
"model.layers.36.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 521 |
+
"model.layers.36.mlp.gate_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 522 |
+
"model.layers.36.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 523 |
+
"model.layers.36.mlp.up_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 524 |
+
"model.layers.36.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 525 |
+
"model.layers.36.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 526 |
+
"model.layers.36.self_attn.k_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 527 |
+
"model.layers.36.self_attn.k_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 528 |
+
"model.layers.36.self_attn.o_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 529 |
+
"model.layers.36.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 530 |
+
"model.layers.36.self_attn.q_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 531 |
+
"model.layers.36.self_attn.q_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 532 |
+
"model.layers.36.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00007.bin",
|
| 533 |
+
"model.layers.36.self_attn.v_proj.SCB": "pytorch_model-00006-of-00007.bin",
|
| 534 |
+
"model.layers.36.self_attn.v_proj.weight": "pytorch_model-00006-of-00007.bin",
|
| 535 |
+
"model.layers.37.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 536 |
+
"model.layers.37.mlp.down_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 537 |
+
"model.layers.37.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 538 |
+
"model.layers.37.mlp.gate_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 539 |
+
"model.layers.37.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 540 |
+
"model.layers.37.mlp.up_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 541 |
+
"model.layers.37.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 542 |
+
"model.layers.37.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 543 |
+
"model.layers.37.self_attn.k_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 544 |
+
"model.layers.37.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 545 |
+
"model.layers.37.self_attn.o_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 546 |
+
"model.layers.37.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 547 |
+
"model.layers.37.self_attn.q_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 548 |
+
"model.layers.37.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 549 |
+
"model.layers.37.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00007.bin",
|
| 550 |
+
"model.layers.37.self_attn.v_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 551 |
+
"model.layers.37.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 552 |
+
"model.layers.38.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 553 |
+
"model.layers.38.mlp.down_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 554 |
+
"model.layers.38.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 555 |
+
"model.layers.38.mlp.gate_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 556 |
+
"model.layers.38.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 557 |
+
"model.layers.38.mlp.up_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 558 |
+
"model.layers.38.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 559 |
+
"model.layers.38.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 560 |
+
"model.layers.38.self_attn.k_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 561 |
+
"model.layers.38.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 562 |
+
"model.layers.38.self_attn.o_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 563 |
+
"model.layers.38.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 564 |
+
"model.layers.38.self_attn.q_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 565 |
+
"model.layers.38.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 566 |
+
"model.layers.38.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00007.bin",
|
| 567 |
+
"model.layers.38.self_attn.v_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 568 |
+
"model.layers.38.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 569 |
+
"model.layers.39.input_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 570 |
+
"model.layers.39.mlp.down_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 571 |
+
"model.layers.39.mlp.down_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 572 |
+
"model.layers.39.mlp.gate_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 573 |
+
"model.layers.39.mlp.gate_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 574 |
+
"model.layers.39.mlp.up_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 575 |
+
"model.layers.39.mlp.up_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 576 |
+
"model.layers.39.post_attention_layernorm.weight": "pytorch_model-00007-of-00007.bin",
|
| 577 |
+
"model.layers.39.self_attn.k_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 578 |
+
"model.layers.39.self_attn.k_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 579 |
+
"model.layers.39.self_attn.o_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 580 |
+
"model.layers.39.self_attn.o_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 581 |
+
"model.layers.39.self_attn.q_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 582 |
+
"model.layers.39.self_attn.q_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 583 |
+
"model.layers.39.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00007.bin",
|
| 584 |
+
"model.layers.39.self_attn.v_proj.SCB": "pytorch_model-00007-of-00007.bin",
|
| 585 |
+
"model.layers.39.self_attn.v_proj.weight": "pytorch_model-00007-of-00007.bin",
|
| 586 |
+
"model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 587 |
+
"model.layers.4.mlp.down_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 588 |
+
"model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 589 |
+
"model.layers.4.mlp.gate_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 590 |
+
"model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 591 |
+
"model.layers.4.mlp.up_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 592 |
+
"model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 593 |
+
"model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00007.bin",
|
| 594 |
+
"model.layers.4.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 595 |
+
"model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 596 |
+
"model.layers.4.self_attn.o_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 597 |
+
"model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 598 |
+
"model.layers.4.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 599 |
+
"model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 600 |
+
"model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00007.bin",
|
| 601 |
+
"model.layers.4.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 602 |
+
"model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 603 |
+
"model.layers.5.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 604 |
+
"model.layers.5.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 605 |
+
"model.layers.5.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 606 |
+
"model.layers.5.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 607 |
+
"model.layers.5.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 608 |
+
"model.layers.5.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 609 |
+
"model.layers.5.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 610 |
+
"model.layers.5.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 611 |
+
"model.layers.5.self_attn.k_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 612 |
+
"model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 613 |
+
"model.layers.5.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 614 |
+
"model.layers.5.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 615 |
+
"model.layers.5.self_attn.q_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 616 |
+
"model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 617 |
+
"model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 618 |
+
"model.layers.5.self_attn.v_proj.SCB": "pytorch_model-00001-of-00007.bin",
|
| 619 |
+
"model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00007.bin",
|
| 620 |
+
"model.layers.6.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 621 |
+
"model.layers.6.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 622 |
+
"model.layers.6.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 623 |
+
"model.layers.6.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 624 |
+
"model.layers.6.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 625 |
+
"model.layers.6.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 626 |
+
"model.layers.6.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 627 |
+
"model.layers.6.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 628 |
+
"model.layers.6.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 629 |
+
"model.layers.6.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 630 |
+
"model.layers.6.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 631 |
+
"model.layers.6.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 632 |
+
"model.layers.6.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 633 |
+
"model.layers.6.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 634 |
+
"model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 635 |
+
"model.layers.6.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 636 |
+
"model.layers.6.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 637 |
+
"model.layers.7.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 638 |
+
"model.layers.7.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 639 |
+
"model.layers.7.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 640 |
+
"model.layers.7.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 641 |
+
"model.layers.7.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 642 |
+
"model.layers.7.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 643 |
+
"model.layers.7.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 644 |
+
"model.layers.7.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 645 |
+
"model.layers.7.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 646 |
+
"model.layers.7.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 647 |
+
"model.layers.7.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 648 |
+
"model.layers.7.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 649 |
+
"model.layers.7.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 650 |
+
"model.layers.7.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 651 |
+
"model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 652 |
+
"model.layers.7.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 653 |
+
"model.layers.7.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 654 |
+
"model.layers.8.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 655 |
+
"model.layers.8.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 656 |
+
"model.layers.8.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 657 |
+
"model.layers.8.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 658 |
+
"model.layers.8.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 659 |
+
"model.layers.8.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 660 |
+
"model.layers.8.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 661 |
+
"model.layers.8.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 662 |
+
"model.layers.8.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 663 |
+
"model.layers.8.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 664 |
+
"model.layers.8.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 665 |
+
"model.layers.8.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 666 |
+
"model.layers.8.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 667 |
+
"model.layers.8.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 668 |
+
"model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 669 |
+
"model.layers.8.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 670 |
+
"model.layers.8.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 671 |
+
"model.layers.9.input_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 672 |
+
"model.layers.9.mlp.down_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 673 |
+
"model.layers.9.mlp.down_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 674 |
+
"model.layers.9.mlp.gate_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 675 |
+
"model.layers.9.mlp.gate_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 676 |
+
"model.layers.9.mlp.up_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 677 |
+
"model.layers.9.mlp.up_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 678 |
+
"model.layers.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00007.bin",
|
| 679 |
+
"model.layers.9.self_attn.k_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 680 |
+
"model.layers.9.self_attn.k_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 681 |
+
"model.layers.9.self_attn.o_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 682 |
+
"model.layers.9.self_attn.o_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 683 |
+
"model.layers.9.self_attn.q_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 684 |
+
"model.layers.9.self_attn.q_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 685 |
+
"model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00007.bin",
|
| 686 |
+
"model.layers.9.self_attn.v_proj.SCB": "pytorch_model-00002-of-00007.bin",
|
| 687 |
+
"model.layers.9.self_attn.v_proj.weight": "pytorch_model-00002-of-00007.bin",
|
| 688 |
+
"model.norm.weight": "pytorch_model-00007-of-00007.bin"
|
| 689 |
+
}
|
| 690 |
+
}
|