Add OpenLLaMA 7B
Browse files- .gitignore +5 -0
- Makefile +47 -0
- README.md +10 -0
- SHA256SUMS +18 -0
- open-llama-7b-f16.bin +3 -0
- open-llama-7b-q2_K.bin +3 -0
- open-llama-7b-q3_K.bin +3 -0
- open-llama-7b-q3_K_L.bin +3 -0
- open-llama-7b-q3_K_M.bin +3 -0
- open-llama-7b-q3_K_S.bin +3 -0
- open-llama-7b-q4_0.bin +3 -0
- open-llama-7b-q4_1.bin +3 -0
- open-llama-7b-q4_K.bin +3 -0
- open-llama-7b-q4_K_M.bin +3 -0
- open-llama-7b-q4_K_S.bin +3 -0
- open-llama-7b-q5_0.bin +3 -0
- open-llama-7b-q5_1.bin +3 -0
- open-llama-7b-q5_K.bin +3 -0
- open-llama-7b-q5_K_M.bin +3 -0
- open-llama-7b-q5_K_S.bin +3 -0
- open-llama-7b-q6_K.bin +3 -0
- open-llama-7b-q8_0.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
llama.cpp/
|
2 |
+
pytorch_model*.bin
|
3 |
+
*.sha
|
4 |
+
*.tar.gz
|
5 |
+
tokenizer.model
|
Makefile
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
MODEL_NAME= open-llama-7b
|
2 |
+
PYTHON?= python
|
3 |
+
LLAMA_TAG= 5c64a09
|
4 |
+
LLAMA_TAR= master-$(LLAMA_TAG).tar.gz
|
5 |
+
HF_REPO= openlm-research/open_llama_7b
|
6 |
+
HF_REF= main
|
7 |
+
HF_FILES= pytorch_model-00001-of-00002.bin \
|
8 |
+
pytorch_model-00002-of-00002.bin \
|
9 |
+
tokenizer.model
|
10 |
+
$(HF_FILES): SITE= https://huggingface.co/$(HF_REPO)/resolve/$(HF_REF)
|
11 |
+
$(LLAMA_TAR): SITE= https://github.com/ggerganov/llama.cpp/archive/refs/tags
|
12 |
+
|
13 |
+
FILES= $(HF_FILES) $(LLAMA_TAR)
|
14 |
+
|
15 |
+
QUANTS= f16 q4_0 q4_1 q5_0 q5_1 q8_0 \
|
16 |
+
q2_K \
|
17 |
+
q3_K q3_K_L q3_K_M q3_K_S \
|
18 |
+
q4_K q4_K_M q4_K_S \
|
19 |
+
q5_K q5_K_M q5_K_S \
|
20 |
+
q6_K
|
21 |
+
|
22 |
+
MODEL_FILES= $(addsuffix .bin,$(addprefix $(MODEL_NAME)-,$(QUANTS)))
|
23 |
+
|
24 |
+
.PHONY: all
|
25 |
+
all: $(MODEL_FILES) SHA256SUMS
|
26 |
+
|
27 |
+
$(FILES):
|
28 |
+
curl -L -o $@ --url $(SITE)/$@
|
29 |
+
|
30 |
+
llama.cpp: $(LLAMA_TAR)
|
31 |
+
mkdir -p $@
|
32 |
+
tar -xf $< --strip-components=1 -C $@
|
33 |
+
|
34 |
+
llama.cpp/quantize: llama.cpp
|
35 |
+
$(MAKE) -C llama.cpp quantize
|
36 |
+
|
37 |
+
$(MODEL_NAME)-f16.bin: $(HF_FILES) | llama.cpp
|
38 |
+
$(PYTHON) llama.cpp/convert.py --outtype f16 --outfile $@ .
|
39 |
+
|
40 |
+
$(MODEL_NAME)-q%.bin: $(MODEL_NAME)-f16.bin | llama.cpp/quantize
|
41 |
+
llama.cpp/quantize $< $@ q$*
|
42 |
+
|
43 |
+
%.sha: %
|
44 |
+
sha256sum $< > $@
|
45 |
+
|
46 |
+
SHA256SUMS: $(addsuffix .sha,$(MODEL_FILES))
|
47 |
+
cat $^ > $@
|
README.md
CHANGED
@@ -1,3 +1,13 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
|
5 |
+
# ggml versions of OpenLLaMa 7B
|
6 |
+
|
7 |
+
For use with [llama.cpp](https://github.com/ggerganov/llama.cpp).
|
8 |
+
|
9 |
+
- Version: 1T tokens final version
|
10 |
+
- Project: [OpenLLaMA: An Open Reproduction of LLaMA](https://github.com/openlm-research/open_llama)
|
11 |
+
- Model: [openlm-research/open_llama_3b](https://huggingface.co/openlm-research/open_llama_7b)
|
12 |
+
- llama.cpp 4,5,8-bit quantization: build 567(2d5db48) or later
|
13 |
+
- llama.cpp newer quantization formats: build 616(99009e7) or later
|
SHA256SUMS
ADDED
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
52eeca58b30e071e22551178d9522e704f04c71f759c9a8f12e3ef99a3d0dc1e open-llama-7b-f16.bin
|
2 |
+
b36fe08e448a2a9c3a254e35f881f904cf197794a71970883bdb562b3eb0e7a4 open-llama-7b-q4_0.bin
|
3 |
+
f0d2ce1d1728c305a1d1c791536c854aa0cfba3b339aba6205c6926983076c1d open-llama-7b-q4_1.bin
|
4 |
+
0214de39ac457203d6ac44189e4523e6f56f99b4072104ee70d088d6c386307b open-llama-7b-q5_0.bin
|
5 |
+
f12cb3e5029a0957b055560f5f1fd1b549662720747c12c67134c0ccf9a43d32 open-llama-7b-q5_1.bin
|
6 |
+
2059d74ad06aab36b226c480fec8bd8fb68d467e66782ae72f4e48ff30cabb17 open-llama-7b-q8_0.bin
|
7 |
+
665bbc8b910965ca006c6e890216b6eab6d162fc9b50b6e22882c524ad135446 open-llama-7b-q2_K.bin
|
8 |
+
b6f0431fbc6fac5126cd5a855545d9fe27f4a18c6168a99ac37c138a4a75b5c4 open-llama-7b-q3_K.bin
|
9 |
+
60c26ed2b0c74232c6aeff1b4096c947dbf69ec0f648abbb9e6d400084b933ae open-llama-7b-q3_K_L.bin
|
10 |
+
b6f0431fbc6fac5126cd5a855545d9fe27f4a18c6168a99ac37c138a4a75b5c4 open-llama-7b-q3_K_M.bin
|
11 |
+
66939ad0975a116d866ab4918132ccb842edbee507ed0af144a53c2a97f5fe3d open-llama-7b-q3_K_S.bin
|
12 |
+
9a48fc57b63afe6218c2c70ec85c998f88a78d14d009928da7a1f4e8643e73a6 open-llama-7b-q4_K.bin
|
13 |
+
9a48fc57b63afe6218c2c70ec85c998f88a78d14d009928da7a1f4e8643e73a6 open-llama-7b-q4_K_M.bin
|
14 |
+
7ab744ce2ffe5bbf8b043179641a0b2dccbe74ff72c58cda717c043c7f6896cf open-llama-7b-q4_K_S.bin
|
15 |
+
233d50a95afd1f4edb7ba856eb5b7e22fa1cd509a80b08291f4f7bb9cf9a0486 open-llama-7b-q5_K.bin
|
16 |
+
233d50a95afd1f4edb7ba856eb5b7e22fa1cd509a80b08291f4f7bb9cf9a0486 open-llama-7b-q5_K_M.bin
|
17 |
+
a8dcb09aef6e57bf1fcd2d2b0eb94878e16241468b2366b862873cbeb5d1f89d open-llama-7b-q5_K_S.bin
|
18 |
+
0e9c8704b7c1c65aa80ac8d758a19b703e40b82ba0d38df4e9d72d1bd3f90485 open-llama-7b-q6_K.bin
|
open-llama-7b-f16.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:52eeca58b30e071e22551178d9522e704f04c71f759c9a8f12e3ef99a3d0dc1e
|
3 |
+
size 13477838976
|
open-llama-7b-q2_K.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:665bbc8b910965ca006c6e890216b6eab6d162fc9b50b6e22882c524ad135446
|
3 |
+
size 2802319488
|
open-llama-7b-q3_K.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b6f0431fbc6fac5126cd5a855545d9fe27f4a18c6168a99ac37c138a4a75b5c4
|
3 |
+
size 3231072384
|
open-llama-7b-q3_K_L.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:60c26ed2b0c74232c6aeff1b4096c947dbf69ec0f648abbb9e6d400084b933ae
|
3 |
+
size 3545645184
|
open-llama-7b-q3_K_M.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b6f0431fbc6fac5126cd5a855545d9fe27f4a18c6168a99ac37c138a4a75b5c4
|
3 |
+
size 3231072384
|
open-llama-7b-q3_K_S.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:66939ad0975a116d866ab4918132ccb842edbee507ed0af144a53c2a97f5fe3d
|
3 |
+
size 2896838784
|
open-llama-7b-q4_0.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b36fe08e448a2a9c3a254e35f881f904cf197794a71970883bdb562b3eb0e7a4
|
3 |
+
size 3791749248
|
open-llama-7b-q4_1.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f0d2ce1d1728c305a1d1c791536c854aa0cfba3b339aba6205c6926983076c1d
|
3 |
+
size 4212883584
|
open-llama-7b-q4_K.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a48fc57b63afe6218c2c70ec85c998f88a78d14d009928da7a1f4e8643e73a6
|
3 |
+
size 4046946432
|
open-llama-7b-q4_K_M.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9a48fc57b63afe6218c2c70ec85c998f88a78d14d009928da7a1f4e8643e73a6
|
3 |
+
size 4046946432
|
open-llama-7b-q4_K_S.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7ab744ce2ffe5bbf8b043179641a0b2dccbe74ff72c58cda717c043c7f6896cf
|
3 |
+
size 3791749248
|
open-llama-7b-q5_0.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0214de39ac457203d6ac44189e4523e6f56f99b4072104ee70d088d6c386307b
|
3 |
+
size 4634017920
|
open-llama-7b-q5_1.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f12cb3e5029a0957b055560f5f1fd1b549662720747c12c67134c0ccf9a43d32
|
3 |
+
size 5055152256
|
open-llama-7b-q5_K.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:233d50a95afd1f4edb7ba856eb5b7e22fa1cd509a80b08291f4f7bb9cf9a0486
|
3 |
+
size 4765483136
|
open-llama-7b-q5_K_M.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:233d50a95afd1f4edb7ba856eb5b7e22fa1cd509a80b08291f4f7bb9cf9a0486
|
3 |
+
size 4765483136
|
open-llama-7b-q5_K_S.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a8dcb09aef6e57bf1fcd2d2b0eb94878e16241468b2366b862873cbeb5d1f89d
|
3 |
+
size 4634017920
|
open-llama-7b-q6_K.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0e9c8704b7c1c65aa80ac8d758a19b703e40b82ba0d38df4e9d72d1bd3f90485
|
3 |
+
size 5528928384
|
open-llama-7b-q8_0.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2059d74ad06aab36b226c480fec8bd8fb68d467e66782ae72f4e48ff30cabb17
|
3 |
+
size 7160823936
|