add files

.gitattributes

@@ -4,6 +4,7 @@
 *.bz2 filter=lfs diff=lfs merge=lfs -text
 *.ckpt filter=lfs diff=lfs merge=lfs -text
 *.ftz filter=lfs diff=lfs merge=lfs -text
+*.gguf filter=lfs diff=lfs merge=lfs -text
 *.gz filter=lfs diff=lfs merge=lfs -text
 *.h5 filter=lfs diff=lfs merge=lfs -text
 *.joblib filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,8 @@
+llama.cpp*/
+venv/
+pytorch_model*.bin
+*.sha
+*.tar.gz
+tokenizer.model
+config.json
+tokenizer_config.json

Makefile

@@ -0,0 +1,54 @@
+MODEL_NAME=	open-llama-7b-v2
+PYTHON?=	python
+LLAMA_BUILD=	1132
+LLAMA_TAR=	b$(LLAMA_BUILD).tar.gz
+LLAMA_DIR=	llama.cpp-b$(LLAMA_BUILD)
+LLAMA_FLAGS=
+HF_REPO=	openlm-research/open_llama_7b_v2
+HF_REF=		main
+HF_FILES=	pytorch_model-00001-of-00002.bin \
+		pytorch_model-00002-of-00002.bin \
+		tokenizer.model \
+		config.json \
+		tokenizer_config.json
+$(HF_FILES): 	SITE=	https://huggingface.co/$(HF_REPO)/resolve/$(HF_REF)
+$(LLAMA_TAR): 	SITE=	https://github.com/ggerganov/llama.cpp/archive/refs/tags
+
+QUANTS=		f16 q4_0 q4_1 q5_0 q5_1 q8_0 \
+		q2_K \
+		q3_K_S q3_K_M q3_K_L \
+		q4_K_S q4_K_M \
+		q5_K_S q5_K_M \
+		q6_K
+
+FILES=		$(HF_FILES) $(LLAMA_TAR)
+MODEL_FILES=	$(foreach q,$(QUANTS),$(MODEL_NAME)-$(q).gguf)
+
+.PHONY: all
+all: $(MODEL_FILES) SHA256SUMS
+
+$(FILES):
+	curl -L -o $@ --url $(SITE)/$@
+
+$(LLAMA_DIR): | $(LLAMA_TAR)
+	tar -xf $(LLAMA_TAR)
+
+$(LLAMA_DIR)/quantize: | $(LLAMA_DIR)
+	$(MAKE) -C $(LLAMA_DIR) $(LLAMA_FLAGS) quantize
+
+venv:
+	$(PYTHON) -m venv venv
+	venv/bin/pip install -e $(LLAMA_DIR)/gguf-py
+	venv/bin/pip install -r $(LLAMA_DIR)/requirements.txt
+
+$(MODEL_NAME)-f16.gguf: $(HF_FILES) | $(LLAMA_DIR) venv
+	venv/bin/python $(LLAMA_DIR)/convert.py --outtype f16 --outfile $@ .
+
+$(MODEL_NAME)-q%.gguf: $(MODEL_NAME)-f16.gguf | $(LLAMA_DIR)/quantize
+	$(LLAMA_DIR)/quantize $< $@ q$*
+
+%.sha: %
+	sha256sum $< > $@
+
+SHA256SUMS: $(addsuffix .sha,$(MODEL_FILES))
+	cat $^ > $@

README.md

@@ -1,3 +1,15 @@
 ---
 license: apache-2.0
 ---
+
+# gguf versions of OpenLLaMa 7B v2
+
+- Version: [version 2 final 1T tokens](https://github.com/openlm-research/open_llama#07072023)
+- Project: [OpenLLaMA: An Open Reproduction of LLaMA](https://github.com/openlm-research/open_llama)
+- Model: [openlm-research/open_llama_7b_v2](https://huggingface.co/openlm-research/open_llama_7b_v2)
+- [llama.cpp](https://github.com/ggerganov/llama.cpp): build 1012 (6381d4e) or later
+- [ggml version](https://huggingface.co/SlyEcho/open_llama_7b_v2_ggml)
+
+## Perplexity on wiki.test.406
+
+Coming soon...

SHA256SUMS

@@ -0,0 +1,15 @@
+79262501cdc3a688ae9f641e37f431d1844934ad28735eff1c6041e7bddc0d8d  open-llama-7b-v2-f16.gguf
+87616025fcf417f857f631db89f0b15238e8d03a621cfc665ab632e8d5e4c789  open-llama-7b-v2-q4_0.gguf
+5a202ba97532d7da0a980b806c1d08b9721325e120842ebb10f2dfff5cf4cb85  open-llama-7b-v2-q4_1.gguf
+9e67e3ca4c34dfe342fe064e600908435b79a163fe17239709943b8b95cb2d07  open-llama-7b-v2-q5_0.gguf
+a2b336edbc93e788a644c79f06565c932637839eb6f078db1726b982c8927962  open-llama-7b-v2-q5_1.gguf
+2aff8fcdff69e26595059ae9252f4cbcfde1e7a7dd246a16c22f1d93b5d65e10  open-llama-7b-v2-q8_0.gguf
+2c22e744d4aac99b85f9607363f4df58e77a89d372053367f900626d8ccfb99b  open-llama-7b-v2-q2_K.gguf
+4d4b76064a6b6a7cc87eb2c335f4e5fe4489411a6e88558dbecf072b0331b56b  open-llama-7b-v2-q3_K_S.gguf
+679b90fb5a54af0e08711279fc4c31d23c06d1b99ee408f03487379fa7901b1f  open-llama-7b-v2-q3_K_M.gguf
+da48217a3997587deffeadeb0a0c573669850bb4ea5b5a147a720b726a9bee5d  open-llama-7b-v2-q3_K_L.gguf
+cf15b3f86bd39ddb20e313129f699e7a6f5b7af90878cd573d5cabff66ca384b  open-llama-7b-v2-q4_K_S.gguf
+5fe444cd2338b3c257ca6ca98c6385e9573989bb75042c5a09de4bf366340d0b  open-llama-7b-v2-q4_K_M.gguf
+4385a7f4f60ab151a808fbba1da1e59963648fa2bc0e38ce1e635ea34c22c881  open-llama-7b-v2-q5_K_S.gguf
+9ca9d74646a1082ee042cd0d05bc12029b8e661fc12a8a651cbadc6fa342b601  open-llama-7b-v2-q5_K_M.gguf
+121553be75551e618757928dc45446efa8d8ea7048a0fff87143c6a1ebbc160e  open-llama-7b-v2-q6_K.gguf

open-llama-7b-v2-f16.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79262501cdc3a688ae9f641e37f431d1844934ad28735eff1c6041e7bddc0d8d
+size 13478116576

open-llama-7b-v2-q2_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c22e744d4aac99b85f9607363f4df58e77a89d372053367f900626d8ccfb99b
+size 2825952544

open-llama-7b-v2-q3_K_L.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da48217a3997587deffeadeb0a0c573669850bb4ea5b5a147a720b726a9bee5d
+size 3597122848

open-llama-7b-v2-q3_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:679b90fb5a54af0e08711279fc4c31d23c06d1b99ee408f03487379fa7901b1f
+size 3298016544

open-llama-7b-v2-q3_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d4b76064a6b6a7cc87eb2c335f4e5fe4489411a6e88558dbecf072b0331b56b
+size 2948316448

open-llama-7b-v2-q4_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87616025fcf417f857f631db89f0b15238e8d03a621cfc665ab632e8d5e4c789
+size 3825818912

open-llama-7b-v2-q4_1.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a202ba97532d7da0a980b806c1d08b9721325e120842ebb10f2dfff5cf4cb85
+size 4238761248

open-llama-7b-v2-q4_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fe444cd2338b3c257ca6ca98c6385e9573989bb75042c5a09de4bf366340d0b
+size 4081016096

open-llama-7b-v2-q4_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf15b3f86bd39ddb20e313129f699e7a6f5b7af90878cd573d5cabff66ca384b
+size 3856751904

open-llama-7b-v2-q5_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e67e3ca4c34dfe342fe064e600908435b79a163fe17239709943b8b95cb2d07
+size 4651703584

open-llama-7b-v2-q5_1.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2b336edbc93e788a644c79f06565c932637839eb6f078db1726b982c8927962
+size 5064645920

open-llama-7b-v2-q5_K_M.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ca9d74646a1082ee042cd0d05bc12029b8e661fc12a8a651cbadc6fa342b601
+size 4783168800

open-llama-7b-v2-q5_K_S.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4385a7f4f60ab151a808fbba1da1e59963648fa2bc0e38ce1e635ea34c22c881
+size 4651703584

open-llama-7b-v2-q6_K.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:121553be75551e618757928dc45446efa8d8ea7048a0fff87143c6a1ebbc160e
+size 5529206048

open-llama-7b-v2-q8_0.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2aff8fcdff69e26595059ae9252f4cbcfde1e7a7dd246a16c22f1d93b5d65e10
+size 7161101600