From 0818415c7f7a5c41f5ba20bb1668a443bb9ebeac Mon Sep 17 00:00:00 2001 From: Yagil Burowski Date: Tue, 31 Oct 2023 20:55:44 -0400 Subject: [PATCH 1/2] update CodeLlama to GGUF --- models/CodeLlama-7B-Instruct.json | 34 ++++++++++++++-------------- models/Mistral-7B-Instruct-v0.1.json | 2 +- models/zephyr-7B-beta.json | 2 +- schema.json | 2 +- 4 files changed, 20 insertions(+), 20 deletions(-) diff --git a/models/CodeLlama-7B-Instruct.json b/models/CodeLlama-7B-Instruct.json index 924cb28..114f8d4 100644 --- a/models/CodeLlama-7B-Instruct.json +++ b/models/CodeLlama-7B-Instruct.json @@ -12,47 +12,47 @@ "resources": { "canonicalUrl": "https://ai.meta.com/blog/code-llama-large-language-model-coding/", "paperUrl": "https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/", - "downloadUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGML" + "downloadUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF" }, "trainedFor": "chat", "arch": "llama", "files": { "highlighted": { "economical": { - "name": "codellama-7b-instruct.ggmlv3.Q4_K_S.bin" + "name": "codellama-7b-instruct.Q4_K_S.gguf" }, "most_capable": { - "name": "codellama-7b-instruct.ggmlv3.Q6_K.bin" + "name": "codellama-7b-instruct.Q6_K.gguf" } }, "all": [ { - "name": "codellama-7b-instruct.ggmlv3.Q4_K_S.bin", - "url": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGML/resolve/main/codellama-7b-instruct.ggmlv3.Q4_K_S.bin", - "sizeBytes": 3980309408, + "name": "codellama-7b-instruct.Q4_K_S.gguf", + "url": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q4_K_S.gguf", + "sizeBytes": 3856831168, "quantization": "Q4_K_S", - "format": "ggml", - "sha256checksum": "b2e7d50d8c854c6654c780d2f3320f83ec2931d5828ebf9777cff00ab7f86fac", + "format": "gguf", + "sha256checksum": "2e44d2b7ae28bbe3a2ed698e259cbd3a6bf7fe8f9d351e14b2be17fb690d7f95", "publisher": { "name": "TheBloke", "socialUrl": "https://twitter.com/TheBlokeAI" }, - "respository": "TheBloke/CodeLlama-7B-Instruct-GGML", - "repositoryUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGML" + "respository": "TheBloke/CodeLlama-7B-Instruct-GGUF", + "repositoryUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF" }, { - "name": "codellama-7b-instruct.ggmlv3.Q6_K.bin", - "url": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGML/resolve/main/codellama-7b-instruct.ggmlv3.Q6_K.bin", - "sizeBytes": 5649904544, + "name": "codellama-7b-instruct.Q6_K.gguf", + "url": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q6_K.gguf", + "sizeBytes": 5529302208, "quantization": "Q6_K", - "format": "ggml", - "sha256checksum": "abdf0ed996933cce5caee27e0be7ec0734add65bd28d7e235a7eab4b3ec3d762", + "format": "gguf", + "sha256checksum": "2f516cd9c16181832ffceaf94b13e8600d88c9bc8d7f75717d25d8c9cf9aa973", "publisher": { "name": "TheBloke", "socialUrl": "https://twitter.com/TheBlokeAI" }, - "respository": "TheBloke/CodeLlama-7B-Instruct-GGML", - "repositoryUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGML" + "respository": "TheBloke/CodeLlama-7B-Instruct-GGUF", + "repositoryUrl": "https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF" } ] } diff --git a/models/Mistral-7B-Instruct-v0.1.json b/models/Mistral-7B-Instruct-v0.1.json index f6adc59..7fa4ae0 100644 --- a/models/Mistral-7B-Instruct-v0.1.json +++ b/models/Mistral-7B-Instruct-v0.1.json @@ -15,7 +15,7 @@ "downloadUrl": "https://huggingface.co/TheBloke/Mistral-7B-Instruct-v0.1-GGUF" }, "trainedFor": "chat", - "arch": "llama", + "arch": "mistral", "files": { "highlighted": { "economical": { diff --git a/models/zephyr-7B-beta.json b/models/zephyr-7B-beta.json index ac2221e..f013024 100644 --- a/models/zephyr-7B-beta.json +++ b/models/zephyr-7B-beta.json @@ -15,7 +15,7 @@ "downloadUrl": "https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF" }, "trainedFor": "chat", - "arch": "llama", + "arch": "mistral", "files": { "highlighted": { "economical": { diff --git a/schema.json b/schema.json index c9653f9..cb0f197 100644 --- a/schema.json +++ b/schema.json @@ -59,7 +59,7 @@ }, "arch": { "type": "string", - "enum": ["llama", "pythia", "gpt_neo_x", "gpt_j", "mpt", "replit", "starcoder", "falcon"] + "enum": ["llama", "pythia", "gpt-neo-x", "gpt-j", "mpt", "replit", "starcoder", "falcon", "mistral"] }, "description": { "type": "string" From 4bb92e2c4aca2c90c230fcd9588f6cf5397a6a64 Mon Sep 17 00:00:00 2001 From: Yagil Burowski Date: Tue, 31 Oct 2023 20:56:15 -0400 Subject: [PATCH 2/2] fix zephyr file size --- models/zephyr-7B-beta.json | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/models/zephyr-7B-beta.json b/models/zephyr-7B-beta.json index f013024..4ccdd9c 100644 --- a/models/zephyr-7B-beta.json +++ b/models/zephyr-7B-beta.json @@ -29,7 +29,7 @@ { "name": "zephyr-7b-beta.Q4_K_S.gguf", "url": "https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/resolve/main/zephyr-7b-beta.Q4_K_S.gguf", - "sizeBytes": 4140373664, + "sizeBytes": 4140373696, "quantization": "Q4_K_S", "format": "gguf", "sha256checksum": "cafa0b85b2efc15ca33023f3b87f8d0c44ddcace16b3fb608280e0eb8f425cb1", @@ -43,7 +43,7 @@ { "name": "zephyr-7b-beta.Q6_K.gguf", "url": "https://huggingface.co/TheBloke/zephyr-7B-beta-GGUF/resolve/main/zephyr-7b-beta.Q6_K.gguf", - "sizeBytes": 5942064800, + "sizeBytes": 5942064832, "quantization": "Q6_K", "format": "gguf", "sha256checksum": "39b52e291eea6040de078283ee5316ff2a317e2b6f59be56724d9b29bada6cfe",