From a035347530869567786f88c28021263df4f94131 Mon Sep 17 00:00:00 2001 From: sanjay920 Date: Thu, 4 Jul 2024 17:36:09 -0700 Subject: [PATCH 1/3] modify functionary model names --- docs/src/components/BenchmarkTable.js | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/src/components/BenchmarkTable.js b/docs/src/components/BenchmarkTable.js index f2ee978..29f1690 100644 --- a/docs/src/components/BenchmarkTable.js +++ b/docs/src/components/BenchmarkTable.js @@ -202,7 +202,7 @@ const data = [ mtBench:'-', }, { - model: 'functionary-medium-v3.0', + model: 'meetkai/functionary-medium-v3.0', params: 70.6, functionCalling: '46.43%', mmlu: '79.85', @@ -212,7 +212,7 @@ const data = [ mtBench:'5.49', }, { - model: 'functionary-small-v2.5', + model: 'meetkai/functionary-small-v2.5', params: 8.03, functionCalling: '57.14%', mmlu: '63.92', From c187cfe955c5e8faa9f606e76689cb19117df314 Mon Sep 17 00:00:00 2001 From: sanjay920 Date: Tue, 9 Jul 2024 00:38:14 -0700 Subject: [PATCH 2/3] update phi-3 gpqa --- docs/docs/models/Phi.md | 2 +- docs/src/components/BenchmarkTable.js | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/docs/models/Phi.md b/docs/docs/models/Phi.md index 4e5c8cd..8fcc8ed 100644 --- a/docs/docs/models/Phi.md +++ b/docs/docs/models/Phi.md @@ -44,7 +44,7 @@ Phi-3 is a state of the art, lightweight model. It performs exceptionally well d Phi-3 Mini 128k Instruct - 68.17 - 30.58 + 25.90 80.44 28.12 7.92 diff --git a/docs/src/components/BenchmarkTable.js b/docs/src/components/BenchmarkTable.js index 29f1690..27198a6 100644 --- a/docs/src/components/BenchmarkTable.js +++ b/docs/src/components/BenchmarkTable.js @@ -118,7 +118,7 @@ const data = [ params: 3.82, functionCalling: '-', mmlu: '68.17', - gpqa: '30.58', + gpqa: '25.90', gsm8k: '80.44', math: '28.12', mtBench: '7.92', From 80284efa4eec5b1186bb33658d403987541928c4 Mon Sep 17 00:00:00 2001 From: sanjay920 Date: Tue, 9 Jul 2024 16:02:14 -0700 Subject: [PATCH 3/3] update Phi numbers --- docs/docs/models/Phi.md | 50 ++++++++++++++------------- docs/src/components/BenchmarkTable.js | 22 ++++++------ 2 files changed, 37 insertions(+), 35 deletions(-) diff --git a/docs/docs/models/Phi.md b/docs/docs/models/Phi.md index 8fcc8ed..bf8b762 100644 --- a/docs/docs/models/Phi.md +++ b/docs/docs/models/Phi.md @@ -43,32 +43,34 @@ Phi-3 is a state of the art, lightweight model. It performs exceptionally well d Phi-3 Mini 128k Instruct - - 68.17 - 25.90 - 80.44 - 28.12 - 7.92 - 51 - 45 - 64 - 0.31875 - 0.28125 - 0.51875 + 69.36 + 27.01 + 83.7 + 32.92 + 8.02 + 21 + 72 + 67 + 0.13125 + 0.45000 + 0.340625 Rubra Enhanced Phi-3 Mini 128k Instruct - 65.71% - 66.66 - 29.24 - 74.09 - 26.84 - 7.45 - 45 - 51 - 64 - 0.28125 - 0.31875 - 0.48125 + 70.0% + 67.87 + 29.69 + 79.45 + 30.80 + 8.21 + 72 + 21 + 67 + 0.45000 + 0.13125 + 0.659375 - \ No newline at end of file + + +* The above is based on the Phi-3 Mini that was updated by Microsoft in June 2024. The original Phi-3 mini came out in April and the Rubra enhanced model has been trained on both versions \ No newline at end of file diff --git a/docs/src/components/BenchmarkTable.js b/docs/src/components/BenchmarkTable.js index ab7d643..846c633 100644 --- a/docs/src/components/BenchmarkTable.js +++ b/docs/src/components/BenchmarkTable.js @@ -117,21 +117,21 @@ const data = [ model: 'Phi-3 Mini 128k Instruct', params: 3.82, functionCalling: '-', - mmlu: '68.17', - gpqa: '25.90', - gsm8k: '80.44', - math: '28.12', - mtBench: '7.92', + mmlu: '69.36', + gpqa: '27.01', + gsm8k: '83.7', + math: '32.92', + mtBench: '8.02', }, { model: 'Rubra Phi-3 Mini 128k Instruct', - params: 4.27, + params: 4.73, functionCalling: '70.00%', - mmlu: '66.66', - gpqa: '29.24', - gsm8k: '74.09', - math: '26.84', - mtBench: '7.45', + mmlu: '67.87', + gpqa: '29.69', + gsm8k: '79.45', + math: '30.80', + mtBench: '8.21', }, { model: 'Qwen2-7B-Instruct',