2024-02-22 08:47:02 Publish script update

app.py

@@ -58,7 +58,7 @@ output_types_df = pd.DataFrame(
 )
 
 summary_df: pd.DataFrame = pd.read_csv("data/2024-02-05 23:33:22.947120_summary.csv")
-time_of_day_comparison_df = pd.read_csv("data/2024-02-06 09:49:19.637072_time_of_day_comparison.csv")
+time_of_day_comparison_df = pd.read_csv("data/time_of_day_comparison.csv")
 general_plots = pd.read_csv("data/general_plots.csv")
 model_costs_df = pd.read_csv("data/2024-02-05 12:03:45.281624_model_costs.csv")
 time_of_day_plots = pd.read_csv("data/time_of_day_plots.csv")

data/2024-02-22 01:20:08.581735_time_of_day_comparison.csv

@@ -0,0 +1,152 @@
+model,time_of_day,execution_time,characters_count,words_count
+01-ai Yi Chat (34B),afternoon,12.255323665086614,512.3645833333334,51.59375
+01-ai Yi Chat (34B),late afternoon,9.417656523840767,384.9583333333333,40.125
+01-ai Yi Chat (34B),evening,8.16779062861488,360.8333333333333,39.375
+01-ai Yi Chat (34B),late evening,9.680033167203268,424.5,40.875
+01-ai Yi Chat (34B),midnight,8.803933773721967,366.8125,39.5625
+Chronos Hermes (13B),afternoon,6.651120054020601,539.5,76.10416666666667
+Chronos Hermes (13B),late afternoon,7.630303209478205,544.625,62.833333333333336
+Chronos Hermes (13B),evening,6.468730966250102,543.05,72.15
+Chronos Hermes (13B),late evening,7.710561646355523,423.1363636363636,46.31818181818182
+Chronos Hermes (13B),midnight,6.555937608083089,438.7857142857143,53.92857142857143
+Chronos Hermes (13B),night,4.168874144554138,360.0,58.5
+Falcon Instruct (7B),afternoon,1.5774729509611387,36.96875,7.020833333333333
+Falcon Instruct (7B),late afternoon,1.0884015900748116,43.833333333333336,8.166666666666666
+Falcon Instruct (7B),evening,0.7874892950057983,39.2,7.4
+Falcon Instruct (7B),late evening,1.2339241071180864,46.791666666666664,8.708333333333334
+Falcon Instruct (7B),midnight,0.9772553077110877,38.5,7.25
+Falcon Instruct (7B),night,1.0781378746032715,42.0,7.5
+LLaMA-2 Chat (7B),afternoon,3.5475496424569024,477.9479166666667,53.135416666666664
+LLaMA-2 Chat (7B),late afternoon,2.7755608558654785,430.45,62.15
+LLaMA-2 Chat (7B),evening,2.9608939344232734,432.2916666666667,62.333333333333336
+LLaMA-2 Chat (7B),late evening,2.869330001913983,452.7083333333333,63.083333333333336
+LLaMA-2 Chat (7B),midnight,3.022005875905355,444.75,63.0
+LLaMA-2 Chat (7B),night,1.686687707901001,377.5,61.5
+Mistral (7B) Instruct v0.2 (Together AI),early morning,3.704360596338908,511.9375,65.41666666666667
+Mistral (7B) Instruct v0.2 (Together AI),morning,3.7677523457274145,501.96153846153845,64.25
+Mistral (7B) Instruct v0.2 (Together AI),afternoon,3.255509059895914,481.76237623762376,60.7970297029703
+Mistral (7B) Instruct v0.2 (Together AI),late afternoon,3.542298033617545,543.5138888888889,65.65277777777777
+Mistral (7B) Instruct v0.2 (Together AI),evening,5.496347131400273,507.48387096774195,63.774193548387096
+Mistral (7B) Instruct v0.2 (Together AI),late evening,3.3933188574654713,515.2794117647059,66.45588235294117
+Mistral (7B) Instruct v0.2 (Together AI),midnight,4.844553523593479,534.296875,67.3125
+Mistral (7B) Instruct v0.2 (Together AI),night,4.0049218302187715,498.2,63.12
+Mistral-7B-Instruct-v0.2,morning,3.734026002883911,362.9,58.35
+Mistral-7B-Instruct-v0.2,afternoon,3.228973722457886,222.0625,32.25
+Mistral-7B-Instruct-v0.2,late afternoon,3.2048643112182615,219.8625,31.8
+Mistral-7B-Instruct-v0.2,evening,3.397640073299408,261.18333333333334,40.1
+Mistral-7B-Instruct-v0.2,late evening,3.389284573495388,175.79375,25.68125
+Mistral-7B-Instruct-v0.2,midnight,1.9149879813194275,37.8,2.95
+Mixtral-8x7B-Instruct-v0.1,early morning,4.526968242530536,285.045,41.86
+Mixtral-8x7B-Instruct-v0.1,morning,3.9661054956285575,304.82,47.28
+Mixtral-8x7B-Instruct-v0.1,afternoon,5.362903979589355,369.3192307692308,54.353846153846156
+Mixtral-8x7B-Instruct-v0.1,late afternoon,5.80184749175942,347.9681818181818,47.27272727272727
+Mixtral-8x7B-Instruct-v0.1,evening,3.6435119574237023,326.69,48.545
+Mixtral-8x7B-Instruct-v0.1,late evening,5.62397656769588,395.15714285714284,49.02857142857143
+Mixtral-8x7B-Instruct-v0.1,midnight,4.639010797279158,323.0394736842105,42.69210526315789
+Mixtral-8x7B-Instruct-v0.1,night,4.009439338194697,301.24545454545455,42.21818181818182
+OpenHermes-2.5-Mistral (7B),afternoon,3.382803752858152,357.8125,32.791666666666664
+OpenHermes-2.5-Mistral (7B),late afternoon,2.52796063820521,281.125,32.125
+OpenHermes-2.5-Mistral (7B),evening,2.7269538966092197,254.08333333333334,26.916666666666668
+OpenHermes-2.5-Mistral (7B),late evening,2.720560759305954,281.125,32.125
+OpenHermes-2.5-Mistral (7B),midnight,2.6601536591847736,258.0625,28.25
+OpenHermes-2.5-Mistral (7B),night,3.330963373184204,179.5,28.0
+Qwen 1.5 Chat (7B),afternoon,2.539412996504042,344.93617021276594,37.37234042553192
+Qwen 1.5 Chat (7B),late afternoon,2.0715979735056558,269.0,30.125
+Qwen 1.5 Chat (7B),evening,2.325377941131592,244.375,27.25
+Qwen 1.5 Chat (7B),late evening,1.9339114997697913,253.66666666666666,27.541666666666668
+Qwen 1.5 Chat (7B),midnight,2.3953611701726913,268.25,30.125
+RedPajama-INCITE Chat (7B),afternoon,1.8110081959854474,57.71875,5.302083333333333
+RedPajama-INCITE Chat (7B),late afternoon,1.7436921298503876,59.625,5.0
+RedPajama-INCITE Chat (7B),evening,1.964497913013805,55.125,4.583333333333333
+RedPajama-INCITE Chat (7B),late evening,2.258105857031686,52.875,4.375
+RedPajama-INCITE Chat (7B),midnight,2.260685852595738,52.875,4.375
+RedPajama-INCITE Chat (7B),night,1.3475391864776611,34.5,5.0
+Snorkel Mistral PairRM DPO (7B),afternoon,34.586263124148054,5710.697916666667,541.0833333333334
+Snorkel Mistral PairRM DPO (7B),late afternoon,22.353231539328892,4714.0,104.375
+Snorkel Mistral PairRM DPO (7B),evening,24.748520737602597,4596.416666666667,92.33333333333333
+Snorkel Mistral PairRM DPO (7B),late evening,24.120955445549704,4623.333333333333,89.625
+Snorkel Mistral PairRM DPO (7B),midnight,23.378775755564373,4646.0,93.3125
+Snorkel Mistral PairRM DPO (7B),night,4.399647831916809,404.5,60.5
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,early morning,2.0273348593711855,372.13,62.53
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,morning,1.9041210174560548,372.05,62.6
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,afternoon,1.8381905496120452,308.795,51.08
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late afternoon,1.7547113946505954,285.17857142857144,46.9
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,evening,1.7984187936782836,313.99,51.96
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late evening,1.5875422928068372,192.78333333333333,31.261111111111113
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,midnight,1.6341248273849487,210.2,34.2
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,night,2.0128010153770446,372.05,62.6
+Vicuna v1.5 (7B),afternoon,3.670576659115878,353.1458333333333,43.791666666666664
+Vicuna v1.5 (7B),late afternoon,3.768600355495106,324.27272727272725,35.77272727272727
+Vicuna v1.5 (7B),evening,3.0816855430603027,296.1,38.55
+Vicuna v1.5 (7B),late evening,3.779375910758972,292.5416666666667,30.125
+Vicuna v1.5 (7B),midnight,3.6318860905511037,294.875,32.5625
+WizardLM v1.2 (13B),afternoon,7.0199061699127885,320.51063829787233,28.43617021276596
+WizardLM v1.2 (13B),late afternoon,4.935352149217025,254.04166666666666,27.833333333333332
+WizardLM v1.2 (13B),evening,4.939989816058766,240.33333333333334,25.791666666666668
+WizardLM v1.2 (13B),late evening,5.11581133660816,226.25,23.0
+WizardLM v1.2 (13B),midnight,5.251699345452445,219.6875,22.1875
+WizardLM v1.2 (13B),night,3.350038528442383,162.0,26.0
+chat-bison (PaLM 2),early morning,2.4665334616388592,381.9214285714286,60.892857142857146
+chat-bison (PaLM 2),morning,2.488477897644043,381.73,60.88
+chat-bison (PaLM 2),afternoon,2.6460144804074215,376.28076923076924,55.965384615384615
+chat-bison (PaLM 2),late afternoon,3.0387172081253744,381.1363636363636,53.35454545454545
+chat-bison (PaLM 2),evening,2.688272579908371,367.07,55.89
+chat-bison (PaLM 2),late evening,2.7250528037548065,382.725,52.95
+chat-bison (PaLM 2),midnight,2.468383938074112,381.95,60.9
+chat-bison (PaLM 2),night,2.460119960308075,381.92,60.94
+chat-bison-32k (PaLM 2 32K),early morning,9.141417106560299,335.75,53.85
+chat-bison-32k (PaLM 2 32K),morning,7.7035503840446475,335.75,53.85
+chat-bison-32k (PaLM 2 32K),afternoon,5.014458654477046,339.3692307692308,49.323076923076925
+chat-bison-32k (PaLM 2 32K),late afternoon,4.5047362284226855,348.65909090909093,48.35454545454545
+chat-bison-32k (PaLM 2 32K),evening,7.7332194912433625,329.46,49.5
+chat-bison-32k (PaLM 2 32K),late evening,7.796841062307358,349.95,47.805
+chat-bison-32k (PaLM 2 32K),midnight,7.7498266498247785,335.75,53.85
+chat-bison-32k (PaLM 2 32K),night,6.491292915344238,335.75,53.85
+gemini-pro,early morning,2.7453590130460435,381.2214285714286,60.964285714285715
+gemini-pro,morning,2.497767536007628,371.93,59.39
+gemini-pro,afternoon,2.816922114008949,360.62307692307695,53.25769230769231
+gemini-pro,late afternoon,2.9268629640903114,364.57272727272726,50.1
+gemini-pro,evening,2.86901999375759,366.4,55.045
+gemini-pro,late evening,3.7189874940246117,390.9142857142857,51.35
+gemini-pro,midnight,3.338477972348531,369.825,55.9125
+gemini-pro,night,2.8375814715210272,374.8,60.04
+gpt-3.5-turbo,early morning,3.787998208734724,403.74444444444447,47.34444444444444
+gpt-3.5-turbo,morning,3.126271222697364,389.9888888888889,50.93888888888889
+gpt-3.5-turbo,afternoon,3.9458200880459375,381.4428571428571,43.76190476190476
+gpt-3.5-turbo,late afternoon,4.384064777692159,396.4357142857143,43.28333333333333
+gpt-3.5-turbo,evening,3.5153889304115657,385.18095238095236,46.01428571428571
+gpt-3.5-turbo,late evening,5.110168156187617,422.73510971786834,45.358934169278996
+gpt-3.5-turbo,midnight,3.822115447632102,426.71923076923076,49.05
+gpt-3.5-turbo,night,6.062970260473398,418.6192307692308,44.926923076923075
+gpt-4,early morning,14.348626694414351,323.5388888888889,40.544444444444444
+gpt-4,morning,12.759107512468733,338.18333333333334,46.85
+gpt-4,afternoon,16.002364798386893,318.3095238095238,38.77142857142857
+gpt-4,late afternoon,16.80607506932254,313.3595238095238,37.49285714285714
+gpt-4,evening,13.841120740345547,318.0809523809524,40.19285714285714
+gpt-4,late evening,14.298642643005493,314.336,37.012
+gpt-4,midnight,12.3578163115329,334.075,41.35
+gpt-4,night,12.813134506115546,316.93461538461537,37.93076923076923
+gpt-4-turbo,early morning,11.555620827939775,357.65555555555557,47.21666666666667
+gpt-4-turbo,morning,13.686854598257277,381.8888888888889,55.02777777777778
+gpt-4-turbo,afternoon,13.997754749229976,351.01190476190476,46.03333333333333
+gpt-4-turbo,late afternoon,22.320911452883767,381.65714285714284,47.35476190476191
+gpt-4-turbo,evening,16.550320884159632,382.31666666666666,48.45
+gpt-4-turbo,late evening,14.592236209392548,413.61,44.8
+gpt-4-turbo,midnight,13.770663784850727,382.7613636363636,47.61818181818182
+gpt-4-turbo,night,14.254795966698573,352.54615384615386,46.37692307692308
+llama-2-70b-chat,early morning,2.8660141522424265,289.6642857142857,44.614285714285714
+llama-2-70b-chat,morning,2.872361832027194,283.35,43.45
+llama-2-70b-chat,afternoon,4.234376892130426,375.44615384615383,55.238461538461536
+llama-2-70b-chat,late afternoon,3.6833307104881365,433.6636363636364,59.445454545454545
+llama-2-70b-chat,evening,2.9706250462084185,317.245,47.4
+llama-2-70b-chat,late evening,4.719581684340602,572.1689655172414,79.83103448275862
+llama-2-70b-chat,midnight,3.249819871626402,346.8875,52.06875
+llama-2-70b-chat,night,2.8264514451677147,313.71,48.42
+zephyr-7b-beta,early morning,3.937663261095683,273.03333333333336,43.45
+zephyr-7b-beta,morning,4.056525647640228,386.3,63.35
+zephyr-7b-beta,afternoon,3.5789777278900146,277.6,44.016666666666666
+zephyr-7b-beta,late afternoon,3.4592524923459447,248.59,38.9
+zephyr-7b-beta,evening,3.5673056403795878,273.3666666666667,43.5
+zephyr-7b-beta,late evening,3.45343524068594,176.23125,27.21875
+zephyr-7b-beta,midnight,3.7253047794103624,217.82083333333333,33.67916666666667
+zephyr-7b-beta,night,3.6954557319482166,216.55833333333334,33.46666666666667

data/2024-02-22 08:41:24.628859_time_of_day_comparison.csv

@@ -0,0 +1,165 @@
+model,time_of_day,execution_time,characters_count,words_count
+01-ai Yi Chat (34B),early morning,8.585494671549116,402.04545454545456,42.90909090909091
+01-ai Yi Chat (34B),afternoon,12.255323665086614,512.3645833333334,51.59375
+01-ai Yi Chat (34B),late afternoon,9.417656523840767,384.9583333333333,40.125
+01-ai Yi Chat (34B),evening,8.16779062861488,360.8333333333333,39.375
+01-ai Yi Chat (34B),late evening,9.680033167203268,424.5,40.875
+01-ai Yi Chat (34B),midnight,9.165155977816195,389.65,39.95
+01-ai Yi Chat (34B),night,9.463179730354472,413.125,40.5625
+Chronos Hermes (13B),early morning,6.895650175639561,458.9736842105263,53.5
+Chronos Hermes (13B),afternoon,6.651120054020601,539.5,76.10416666666667
+Chronos Hermes (13B),late afternoon,7.630303209478205,544.625,62.833333333333336
+Chronos Hermes (13B),evening,6.468730966250102,543.05,72.15
+Chronos Hermes (13B),late evening,7.710561646355523,423.1363636363636,46.31818181818182
+Chronos Hermes (13B),midnight,7.482297485524958,501.55263157894734,58.89473684210526
+Chronos Hermes (13B),night,7.764214697091476,546.28,62.56
+Falcon Instruct (7B),early morning,1.2506394997621193,47.97727272727273,8.818181818181818
+Falcon Instruct (7B),afternoon,1.5774729509611387,36.96875,7.020833333333333
+Falcon Instruct (7B),late afternoon,1.0884015900748116,43.833333333333336,8.166666666666666
+Falcon Instruct (7B),evening,0.7874892950057983,39.2,7.4
+Falcon Instruct (7B),late evening,1.2339241071180864,46.791666666666664,8.708333333333334
+Falcon Instruct (7B),midnight,0.9522609934210777,37.45,7.075
+Falcon Instruct (7B),night,1.4049408435821533,50.6,9.32
+LLaMA-2 Chat (7B),early morning,2.838810222489493,438.5681818181818,62.68181818181818
+LLaMA-2 Chat (7B),afternoon,3.5475496424569024,477.9479166666667,53.135416666666664
+LLaMA-2 Chat (7B),late afternoon,2.7755608558654785,430.45,62.15
+LLaMA-2 Chat (7B),evening,2.9608939344232734,432.2916666666667,62.333333333333336
+LLaMA-2 Chat (7B),late evening,2.869330001913983,452.7083333333333,63.083333333333336
+LLaMA-2 Chat (7B),midnight,2.8979568664844217,459.075,63.15
+LLaMA-2 Chat (7B),night,6.160623960196972,424.70588235294116,57.029411764705884
+Mistral (7B) Instruct v0.2 (Together AI),early morning,3.5434710951412427,501.5652173913044,64.05434782608695
+Mistral (7B) Instruct v0.2 (Together AI),morning,3.7677523457274145,501.96153846153845,64.25
+Mistral (7B) Instruct v0.2 (Together AI),afternoon,3.255509059895914,481.76237623762376,60.7970297029703
+Mistral (7B) Instruct v0.2 (Together AI),late afternoon,3.542298033617545,543.5138888888889,65.65277777777777
+Mistral (7B) Instruct v0.2 (Together AI),evening,5.496347131400273,507.48387096774195,63.774193548387096
+Mistral (7B) Instruct v0.2 (Together AI),late evening,3.3933188574654713,515.2794117647059,66.45588235294117
+Mistral (7B) Instruct v0.2 (Together AI),midnight,4.5804917102768306,519.4659090909091,65.63636363636364
+Mistral (7B) Instruct v0.2 (Together AI),night,3.621489847860029,513.0204081632653,64.51020408163265
+Mistral-7B-Instruct-v0.2,morning,3.734026002883911,362.9,58.35
+Mistral-7B-Instruct-v0.2,afternoon,3.228973722457886,222.0625,32.25
+Mistral-7B-Instruct-v0.2,late afternoon,3.2048643112182615,219.8625,31.8
+Mistral-7B-Instruct-v0.2,evening,3.397640073299408,261.18333333333334,40.1
+Mistral-7B-Instruct-v0.2,late evening,3.389284573495388,175.79375,25.68125
+Mistral-7B-Instruct-v0.2,midnight,1.9149879813194275,37.8,2.95
+Mixtral-8x7B-Instruct-v0.1,early morning,4.526968242530536,285.045,41.86
+Mixtral-8x7B-Instruct-v0.1,morning,3.9661054956285575,304.82,47.28
+Mixtral-8x7B-Instruct-v0.1,afternoon,5.362903979589355,369.3192307692308,54.353846153846156
+Mixtral-8x7B-Instruct-v0.1,late afternoon,5.80184749175942,347.9681818181818,47.27272727272727
+Mixtral-8x7B-Instruct-v0.1,evening,3.6435119574237023,326.69,48.545
+Mixtral-8x7B-Instruct-v0.1,late evening,5.62397656769588,395.15714285714284,49.02857142857143
+Mixtral-8x7B-Instruct-v0.1,midnight,4.639010797279158,323.0394736842105,42.69210526315789
+Mixtral-8x7B-Instruct-v0.1,night,4.009439338194697,301.24545454545455,42.21818181818182
+OpenHermes-2.5-Mistral (7B),early morning,2.60847454179417,288.25,34.68181818181818
+OpenHermes-2.5-Mistral (7B),afternoon,3.382803752858152,357.8125,32.791666666666664
+OpenHermes-2.5-Mistral (7B),late afternoon,2.52796063820521,281.125,32.125
+OpenHermes-2.5-Mistral (7B),evening,2.7269538966092197,254.08333333333334,26.916666666666668
+OpenHermes-2.5-Mistral (7B),late evening,2.720560759305954,281.125,32.125
+OpenHermes-2.5-Mistral (7B),midnight,2.6603748981769266,271.975,30.575
+OpenHermes-2.5-Mistral (7B),night,2.5896553993225098,277.06,31.96
+Qwen 1.5 Chat (7B),early morning,2.2849691672758623,275.72727272727275,32.63636363636363
+Qwen 1.5 Chat (7B),afternoon,2.539412996504042,344.93617021276594,37.37234042553192
+Qwen 1.5 Chat (7B),late afternoon,2.0715979735056558,269.0,30.125
+Qwen 1.5 Chat (7B),evening,2.325377941131592,244.375,27.25
+Qwen 1.5 Chat (7B),late evening,1.9339114997697913,253.66666666666666,27.541666666666668
+Qwen 1.5 Chat (7B),midnight,2.3441089570522307,268.7,30.125
+Qwen 1.5 Chat (7B),night,1.8622158120075862,269.0,30.125
+RedPajama-INCITE Chat (7B),early morning,2.1242939035097756,51.642857142857146,4.285714285714286
+RedPajama-INCITE Chat (7B),afternoon,1.8110081959854474,57.71875,5.302083333333333
+RedPajama-INCITE Chat (7B),late afternoon,1.7436921298503876,59.625,5.0
+RedPajama-INCITE Chat (7B),evening,1.964497913013805,55.125,4.583333333333333
+RedPajama-INCITE Chat (7B),late evening,2.258105857031686,52.875,4.375
+RedPajama-INCITE Chat (7B),midnight,2.118554650126277,55.575,4.625
+RedPajama-INCITE Chat (7B),night,1.8566838451053784,55.38,4.7
+Snorkel Mistral PairRM DPO (7B),early morning,20.92480931171151,4325.272727272727,101.11363636363636
+Snorkel Mistral PairRM DPO (7B),afternoon,34.586263124148054,5710.697916666667,541.0833333333334
+Snorkel Mistral PairRM DPO (7B),late afternoon,22.353231539328892,4714.0,104.375
+Snorkel Mistral PairRM DPO (7B),evening,24.748520737602597,4596.416666666667,92.33333333333333
+Snorkel Mistral PairRM DPO (7B),late evening,24.120955445549704,4623.333333333333,89.625
+Snorkel Mistral PairRM DPO (7B),midnight,22.80718225087875,4686.8,99.95
+Snorkel Mistral PairRM DPO (7B),night,22.170101329684258,4498.1,95.54
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,early morning,2.0273348593711855,372.13,62.53
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,morning,1.9041210174560548,372.05,62.6
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,afternoon,1.8381905496120452,308.795,51.08
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late afternoon,1.7547113946505954,285.17857142857144,46.9
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,evening,1.7984187936782836,313.99,51.96
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late evening,1.5875422928068372,192.78333333333333,31.261111111111113
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,midnight,1.6341248273849487,210.2,34.2
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,night,2.0128010153770446,372.05,62.6
+Vicuna v1.5 (7B),early morning,3.1717429594560103,262.23809523809524,32.45238095238095
+Vicuna v1.5 (7B),afternoon,3.670576659115878,353.1458333333333,43.791666666666664
+Vicuna v1.5 (7B),late afternoon,3.768600355495106,324.27272727272725,35.77272727272727
+Vicuna v1.5 (7B),evening,3.0816855430603027,296.1,38.55
+Vicuna v1.5 (7B),late evening,3.779375910758972,292.5416666666667,30.125
+Vicuna v1.5 (7B),midnight,3.760787125996181,290.125,32.525
+Vicuna v1.5 (7B),night,3.626542616974224,296.0833333333333,31.458333333333332
+WizardLM v1.2 (13B),early morning,5.036556505021595,253.8181818181818,29.386363636363637
+WizardLM v1.2 (13B),afternoon,7.0199061699127885,320.51063829787233,28.43617021276596
+WizardLM v1.2 (13B),late afternoon,4.935352149217025,254.04166666666666,27.833333333333332
+WizardLM v1.2 (13B),evening,4.939989816058766,240.33333333333334,25.791666666666668
+WizardLM v1.2 (13B),late evening,5.11581133660816,226.25,23.0
+WizardLM v1.2 (13B),midnight,5.1245915309802905,239.825,25.5
+WizardLM v1.2 (13B),night,5.04353141784668,253.16,28.28
+chat-bison (PaLM 2),early morning,2.4665334616388592,381.9214285714286,60.892857142857146
+chat-bison (PaLM 2),morning,2.488477897644043,381.73,60.88
+chat-bison (PaLM 2),afternoon,2.6460144804074215,376.28076923076924,55.965384615384615
+chat-bison (PaLM 2),late afternoon,3.0387172081253744,381.1363636363636,53.35454545454545
+chat-bison (PaLM 2),evening,2.688272579908371,367.07,55.89
+chat-bison (PaLM 2),late evening,2.7250528037548065,382.725,52.95
+chat-bison (PaLM 2),midnight,2.468383938074112,381.95,60.9
+chat-bison (PaLM 2),night,2.460119960308075,381.92,60.94
+chat-bison-32k (PaLM 2 32K),early morning,9.141417106560299,335.75,53.85
+chat-bison-32k (PaLM 2 32K),morning,7.7035503840446475,335.75,53.85
+chat-bison-32k (PaLM 2 32K),afternoon,5.014458654477046,339.3692307692308,49.323076923076925
+chat-bison-32k (PaLM 2 32K),late afternoon,4.5047362284226855,348.65909090909093,48.35454545454545
+chat-bison-32k (PaLM 2 32K),evening,7.7332194912433625,329.46,49.5
+chat-bison-32k (PaLM 2 32K),late evening,7.796841062307358,349.95,47.805
+chat-bison-32k (PaLM 2 32K),midnight,7.7498266498247785,335.75,53.85
+chat-bison-32k (PaLM 2 32K),night,6.491292915344238,335.75,53.85
+gemini-pro,early morning,2.7453590130460435,381.2214285714286,60.964285714285715
+gemini-pro,morning,2.497767536007628,371.93,59.39
+gemini-pro,afternoon,2.816922114008949,360.62307692307695,53.25769230769231
+gemini-pro,late afternoon,2.9268629640903114,364.57272727272726,50.1
+gemini-pro,evening,2.86901999375759,366.4,55.045
+gemini-pro,late evening,3.7189874940246117,390.9142857142857,51.35
+gemini-pro,midnight,3.338477972348531,369.825,55.9125
+gemini-pro,night,2.8375814715210272,374.8,60.04
+gpt-3.5-turbo,early morning,3.787998208734724,403.74444444444447,47.34444444444444
+gpt-3.5-turbo,morning,3.126271222697364,389.9888888888889,50.93888888888889
+gpt-3.5-turbo,afternoon,3.9458200880459375,381.4428571428571,43.76190476190476
+gpt-3.5-turbo,late afternoon,4.384064777692159,396.4357142857143,43.28333333333333
+gpt-3.5-turbo,evening,3.5153889304115657,385.18095238095236,46.01428571428571
+gpt-3.5-turbo,late evening,5.110168156187617,422.73510971786834,45.358934169278996
+gpt-3.5-turbo,midnight,3.822115447632102,426.71923076923076,49.05
+gpt-3.5-turbo,night,6.062970260473398,418.6192307692308,44.926923076923075
+gpt-4,early morning,14.348626694414351,323.5388888888889,40.544444444444444
+gpt-4,morning,12.759107512468733,338.18333333333334,46.85
+gpt-4,afternoon,16.002364798386893,318.3095238095238,38.77142857142857
+gpt-4,late afternoon,16.80607506932254,313.3595238095238,37.49285714285714
+gpt-4,evening,13.841120740345547,318.0809523809524,40.19285714285714
+gpt-4,late evening,14.298642643005493,314.336,37.012
+gpt-4,midnight,12.3578163115329,334.075,41.35
+gpt-4,night,12.813134506115546,316.93461538461537,37.93076923076923
+gpt-4-turbo,early morning,11.555620827939775,357.65555555555557,47.21666666666667
+gpt-4-turbo,morning,13.686854598257277,381.8888888888889,55.02777777777778
+gpt-4-turbo,afternoon,13.997754749229976,351.01190476190476,46.03333333333333
+gpt-4-turbo,late afternoon,22.320911452883767,381.65714285714284,47.35476190476191
+gpt-4-turbo,evening,16.550320884159632,382.31666666666666,48.45
+gpt-4-turbo,late evening,14.592236209392548,413.61,44.8
+gpt-4-turbo,midnight,13.770663784850727,382.7613636363636,47.61818181818182
+gpt-4-turbo,night,14.254795966698573,352.54615384615386,46.37692307692308
+llama-2-70b-chat,early morning,2.8660141522424265,289.6642857142857,44.614285714285714
+llama-2-70b-chat,morning,2.872361832027194,283.35,43.45
+llama-2-70b-chat,afternoon,4.234376892130426,375.44615384615383,55.238461538461536
+llama-2-70b-chat,late afternoon,3.6833307104881365,433.6636363636364,59.445454545454545
+llama-2-70b-chat,evening,2.9706250462084185,317.245,47.4
+llama-2-70b-chat,late evening,4.719581684340602,572.1689655172414,79.83103448275862
+llama-2-70b-chat,midnight,3.249819871626402,346.8875,52.06875
+llama-2-70b-chat,night,2.8264514451677147,313.71,48.42
+zephyr-7b-beta,early morning,3.937663261095683,273.03333333333336,43.45
+zephyr-7b-beta,morning,4.056525647640228,386.3,63.35
+zephyr-7b-beta,afternoon,3.5789777278900146,277.6,44.016666666666666
+zephyr-7b-beta,late afternoon,3.4592524923459447,248.59,38.9
+zephyr-7b-beta,evening,3.5673056403795878,273.3666666666667,43.5
+zephyr-7b-beta,late evening,3.45343524068594,176.23125,27.21875
+zephyr-7b-beta,midnight,3.7253047794103624,217.82083333333333,33.67916666666667
+zephyr-7b-beta,night,3.6954557319482166,216.55833333333334,33.46666666666667

data/time_of_day_comparison.csv

@@ -0,0 +1,165 @@
+model,time_of_day,execution_time,characters_count,words_count
+01-ai Yi Chat (34B),early morning,8.896911223729452,394.1666666666667,40.125
+01-ai Yi Chat (34B),afternoon,12.255323665086614,512.3645833333334,51.59375
+01-ai Yi Chat (34B),late afternoon,9.417656523840767,384.9583333333333,40.125
+01-ai Yi Chat (34B),evening,8.16779062861488,360.8333333333333,39.375
+01-ai Yi Chat (34B),late evening,9.680033167203268,424.5,40.875
+01-ai Yi Chat (34B),midnight,9.165155977816195,389.65,39.95
+01-ai Yi Chat (34B),night,9.463179730354472,413.125,40.5625
+Chronos Hermes (13B),early morning,7.70508497012289,461.4761904761905,49.595238095238095
+Chronos Hermes (13B),afternoon,6.651120054020601,539.5,76.10416666666667
+Chronos Hermes (13B),late afternoon,7.630303209478205,544.625,62.833333333333336
+Chronos Hermes (13B),evening,6.468730966250102,543.05,72.15
+Chronos Hermes (13B),late evening,7.710561646355523,423.1363636363636,46.31818181818182
+Chronos Hermes (13B),midnight,7.482297485524958,501.55263157894734,58.89473684210526
+Chronos Hermes (13B),night,7.764214697091476,546.28,62.56
+Falcon Instruct (7B),early morning,1.1807806547297988,45.6875,8.4375
+Falcon Instruct (7B),afternoon,1.5774729509611387,36.96875,7.020833333333333
+Falcon Instruct (7B),late afternoon,1.0884015900748116,43.833333333333336,8.166666666666666
+Falcon Instruct (7B),evening,0.7874892950057983,39.2,7.4
+Falcon Instruct (7B),late evening,1.2339241071180864,46.791666666666664,8.708333333333334
+Falcon Instruct (7B),midnight,0.9522609934210777,37.45,7.075
+Falcon Instruct (7B),night,1.4049408435821533,50.6,9.32
+LLaMA-2 Chat (7B),early morning,2.8788051708884863,451.5833333333333,62.9375
+LLaMA-2 Chat (7B),afternoon,3.5475496424569024,477.9479166666667,53.135416666666664
+LLaMA-2 Chat (7B),late afternoon,2.7755608558654785,430.45,62.15
+LLaMA-2 Chat (7B),evening,2.9608939344232734,432.2916666666667,62.333333333333336
+LLaMA-2 Chat (7B),late evening,2.869330001913983,452.7083333333333,63.083333333333336
+LLaMA-2 Chat (7B),midnight,2.8979568664844217,459.075,63.15
+LLaMA-2 Chat (7B),night,6.160623960196972,424.70588235294116,57.029411764705884
+Mistral (7B) Instruct v0.2 (Together AI),early morning,3.611201712254728,509.75,64.39583333333333
+Mistral (7B) Instruct v0.2 (Together AI),morning,3.7677523457274145,501.96153846153845,64.25
+Mistral (7B) Instruct v0.2 (Together AI),afternoon,3.255509059895914,481.76237623762376,60.7970297029703
+Mistral (7B) Instruct v0.2 (Together AI),late afternoon,3.542298033617545,543.5138888888889,65.65277777777777
+Mistral (7B) Instruct v0.2 (Together AI),evening,5.496347131400273,507.48387096774195,63.774193548387096
+Mistral (7B) Instruct v0.2 (Together AI),late evening,3.3933188574654713,515.2794117647059,66.45588235294117
+Mistral (7B) Instruct v0.2 (Together AI),midnight,4.5804917102768306,519.4659090909091,65.63636363636364
+Mistral (7B) Instruct v0.2 (Together AI),night,3.621489847860029,513.0204081632653,64.51020408163265
+Mistral-7B-Instruct-v0.2,morning,3.734026002883911,362.9,58.35
+Mistral-7B-Instruct-v0.2,afternoon,3.228973722457886,222.0625,32.25
+Mistral-7B-Instruct-v0.2,late afternoon,3.2048643112182615,219.8625,31.8
+Mistral-7B-Instruct-v0.2,evening,3.397640073299408,261.18333333333334,40.1
+Mistral-7B-Instruct-v0.2,late evening,3.389284573495388,175.79375,25.68125
+Mistral-7B-Instruct-v0.2,midnight,1.9149879813194275,37.8,2.95
+Mixtral-8x7B-Instruct-v0.1,early morning,4.526968242530536,285.045,41.86
+Mixtral-8x7B-Instruct-v0.1,morning,3.9661054956285575,304.82,47.28
+Mixtral-8x7B-Instruct-v0.1,afternoon,5.362903979589355,369.3192307692308,54.353846153846156
+Mixtral-8x7B-Instruct-v0.1,late afternoon,5.80184749175942,347.9681818181818,47.27272727272727
+Mixtral-8x7B-Instruct-v0.1,evening,3.6435119574237023,326.69,48.545
+Mixtral-8x7B-Instruct-v0.1,late evening,5.62397656769588,395.15714285714284,49.02857142857143
+Mixtral-8x7B-Instruct-v0.1,midnight,4.639010797279158,323.0394736842105,42.69210526315789
+Mixtral-8x7B-Instruct-v0.1,night,4.009439338194697,301.24545454545455,42.21818181818182
+OpenHermes-2.5-Mistral (7B),early morning,2.613155171275139,281.125,32.125
+OpenHermes-2.5-Mistral (7B),afternoon,3.382803752858152,357.8125,32.791666666666664
+OpenHermes-2.5-Mistral (7B),late afternoon,2.52796063820521,281.125,32.125
+OpenHermes-2.5-Mistral (7B),evening,2.7269538966092197,254.08333333333334,26.916666666666668
+OpenHermes-2.5-Mistral (7B),late evening,2.720560759305954,281.125,32.125
+OpenHermes-2.5-Mistral (7B),midnight,2.6603748981769266,271.975,30.575
+OpenHermes-2.5-Mistral (7B),night,2.5896553993225098,277.06,31.96
+Qwen 1.5 Chat (7B),early morning,2.234119971593221,269.0625,30.125
+Qwen 1.5 Chat (7B),afternoon,2.539412996504042,344.93617021276594,37.37234042553192
+Qwen 1.5 Chat (7B),late afternoon,2.0715979735056558,269.0,30.125
+Qwen 1.5 Chat (7B),evening,2.325377941131592,244.375,27.25
+Qwen 1.5 Chat (7B),late evening,1.9339114997697913,253.66666666666666,27.541666666666668
+Qwen 1.5 Chat (7B),midnight,2.3441089570522307,268.7,30.125
+Qwen 1.5 Chat (7B),night,1.8622158120075862,269.0,30.125
+RedPajama-INCITE Chat (7B),early morning,2.1123720943927764,52.82608695652174,4.3478260869565215
+RedPajama-INCITE Chat (7B),afternoon,1.8110081959854474,57.71875,5.302083333333333
+RedPajama-INCITE Chat (7B),late afternoon,1.7436921298503876,59.625,5.0
+RedPajama-INCITE Chat (7B),evening,1.964497913013805,55.125,4.583333333333333
+RedPajama-INCITE Chat (7B),late evening,2.258105857031686,52.875,4.375
+RedPajama-INCITE Chat (7B),midnight,2.118554650126277,55.575,4.625
+RedPajama-INCITE Chat (7B),night,1.8566838451053784,55.38,4.7
+Snorkel Mistral PairRM DPO (7B),early morning,22.729273674335886,4691.333333333333,100.6875
+Snorkel Mistral PairRM DPO (7B),afternoon,34.586263124148054,5710.697916666667,541.0833333333334
+Snorkel Mistral PairRM DPO (7B),late afternoon,22.353231539328892,4714.0,104.375
+Snorkel Mistral PairRM DPO (7B),evening,24.748520737602597,4596.416666666667,92.33333333333333
+Snorkel Mistral PairRM DPO (7B),late evening,24.120955445549704,4623.333333333333,89.625
+Snorkel Mistral PairRM DPO (7B),midnight,22.80718225087875,4686.8,99.95
+Snorkel Mistral PairRM DPO (7B),night,22.170101329684258,4498.1,95.54
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,early morning,2.0273348593711855,372.13,62.53
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,morning,1.9041210174560548,372.05,62.6
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,afternoon,1.8381905496120452,308.795,51.08
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late afternoon,1.7547113946505954,285.17857142857144,46.9
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,evening,1.7984187936782836,313.99,51.96
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,late evening,1.5875422928068372,192.78333333333333,31.261111111111113
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,midnight,1.6341248273849487,210.2,34.2
+TinyLlama/TinyLlama-1.1B-Chat-v1.0,night,2.0128010153770446,372.05,62.6
+Vicuna v1.5 (7B),early morning,3.347896182859266,263.45652173913044,29.847826086956523
+Vicuna v1.5 (7B),afternoon,3.670576659115878,353.1458333333333,43.791666666666664
+Vicuna v1.5 (7B),late afternoon,3.768600355495106,324.27272727272725,35.77272727272727
+Vicuna v1.5 (7B),evening,3.0816855430603027,296.1,38.55
+Vicuna v1.5 (7B),late evening,3.779375910758972,292.5416666666667,30.125
+Vicuna v1.5 (7B),midnight,3.760787125996181,290.125,32.525
+Vicuna v1.5 (7B),night,3.626542616974224,296.0833333333333,31.458333333333332
+WizardLM v1.2 (13B),early morning,5.147532618564108,250.89583333333334,27.416666666666668
+WizardLM v1.2 (13B),afternoon,7.0199061699127885,320.51063829787233,28.43617021276596
+WizardLM v1.2 (13B),late afternoon,4.935352149217025,254.04166666666666,27.833333333333332
+WizardLM v1.2 (13B),evening,4.939989816058766,240.33333333333334,25.791666666666668
+WizardLM v1.2 (13B),late evening,5.11581133660816,226.25,23.0
+WizardLM v1.2 (13B),midnight,5.1245915309802905,239.825,25.5
+WizardLM v1.2 (13B),night,5.04353141784668,253.16,28.28
+chat-bison (PaLM 2),early morning,2.4665334616388592,381.9214285714286,60.892857142857146
+chat-bison (PaLM 2),morning,2.488477897644043,381.73,60.88
+chat-bison (PaLM 2),afternoon,2.6460144804074215,376.28076923076924,55.965384615384615
+chat-bison (PaLM 2),late afternoon,3.0387172081253744,381.1363636363636,53.35454545454545
+chat-bison (PaLM 2),evening,2.688272579908371,367.07,55.89
+chat-bison (PaLM 2),late evening,2.7250528037548065,382.725,52.95
+chat-bison (PaLM 2),midnight,2.468383938074112,381.95,60.9
+chat-bison (PaLM 2),night,2.460119960308075,381.92,60.94
+chat-bison-32k (PaLM 2 32K),early morning,9.141417106560299,335.75,53.85
+chat-bison-32k (PaLM 2 32K),morning,7.7035503840446475,335.75,53.85
+chat-bison-32k (PaLM 2 32K),afternoon,5.014458654477046,339.3692307692308,49.323076923076925
+chat-bison-32k (PaLM 2 32K),late afternoon,4.5047362284226855,348.65909090909093,48.35454545454545
+chat-bison-32k (PaLM 2 32K),evening,7.7332194912433625,329.46,49.5
+chat-bison-32k (PaLM 2 32K),late evening,7.796841062307358,349.95,47.805
+chat-bison-32k (PaLM 2 32K),midnight,7.7498266498247785,335.75,53.85
+chat-bison-32k (PaLM 2 32K),night,6.491292915344238,335.75,53.85
+gemini-pro,early morning,2.7453590130460435,381.2214285714286,60.964285714285715
+gemini-pro,morning,2.497767536007628,371.93,59.39
+gemini-pro,afternoon,2.816922114008949,360.62307692307695,53.25769230769231
+gemini-pro,late afternoon,2.9268629640903114,364.57272727272726,50.1
+gemini-pro,evening,2.86901999375759,366.4,55.045
+gemini-pro,late evening,3.7189874940246117,390.9142857142857,51.35
+gemini-pro,midnight,3.338477972348531,369.825,55.9125
+gemini-pro,night,2.8375814715210272,374.8,60.04
+gpt-3.5-turbo,early morning,3.787998208734724,403.74444444444447,47.34444444444444
+gpt-3.5-turbo,morning,3.126271222697364,389.9888888888889,50.93888888888889
+gpt-3.5-turbo,afternoon,3.9458200880459375,381.4428571428571,43.76190476190476
+gpt-3.5-turbo,late afternoon,4.384064777692159,396.4357142857143,43.28333333333333
+gpt-3.5-turbo,evening,3.5153889304115657,385.18095238095236,46.01428571428571
+gpt-3.5-turbo,late evening,5.110168156187617,422.73510971786834,45.358934169278996
+gpt-3.5-turbo,midnight,3.822115447632102,426.71923076923076,49.05
+gpt-3.5-turbo,night,6.062970260473398,418.6192307692308,44.926923076923075
+gpt-4,early morning,14.348626694414351,323.5388888888889,40.544444444444444
+gpt-4,morning,12.759107512468733,338.18333333333334,46.85
+gpt-4,afternoon,16.002364798386893,318.3095238095238,38.77142857142857
+gpt-4,late afternoon,16.80607506932254,313.3595238095238,37.49285714285714
+gpt-4,evening,13.841120740345547,318.0809523809524,40.19285714285714
+gpt-4,late evening,14.298642643005493,314.336,37.012
+gpt-4,midnight,12.3578163115329,334.075,41.35
+gpt-4,night,12.813134506115546,316.93461538461537,37.93076923076923
+gpt-4-turbo,early morning,11.555620827939775,357.65555555555557,47.21666666666667
+gpt-4-turbo,morning,13.686854598257277,381.8888888888889,55.02777777777778
+gpt-4-turbo,afternoon,13.997754749229976,351.01190476190476,46.03333333333333
+gpt-4-turbo,late afternoon,22.320911452883767,381.65714285714284,47.35476190476191
+gpt-4-turbo,evening,16.550320884159632,382.31666666666666,48.45
+gpt-4-turbo,late evening,14.592236209392548,413.61,44.8
+gpt-4-turbo,midnight,13.770663784850727,382.7613636363636,47.61818181818182
+gpt-4-turbo,night,14.254795966698573,352.54615384615386,46.37692307692308
+llama-2-70b-chat,early morning,2.8660141522424265,289.6642857142857,44.614285714285714
+llama-2-70b-chat,morning,2.872361832027194,283.35,43.45
+llama-2-70b-chat,afternoon,4.234376892130426,375.44615384615383,55.238461538461536
+llama-2-70b-chat,late afternoon,3.6833307104881365,433.6636363636364,59.445454545454545
+llama-2-70b-chat,evening,2.9706250462084185,317.245,47.4
+llama-2-70b-chat,late evening,4.719581684340602,572.1689655172414,79.83103448275862
+llama-2-70b-chat,midnight,3.249819871626402,346.8875,52.06875
+llama-2-70b-chat,night,2.8264514451677147,313.71,48.42
+zephyr-7b-beta,early morning,3.937663261095683,273.03333333333336,43.45
+zephyr-7b-beta,morning,4.056525647640228,386.3,63.35
+zephyr-7b-beta,afternoon,3.5789777278900146,277.6,44.016666666666666
+zephyr-7b-beta,late afternoon,3.4592524923459447,248.59,38.9
+zephyr-7b-beta,evening,3.5673056403795878,273.3666666666667,43.5
+zephyr-7b-beta,late evening,3.45343524068594,176.23125,27.21875
+zephyr-7b-beta,midnight,3.7253047794103624,217.82083333333333,33.67916666666667
+zephyr-7b-beta,night,3.6954557319482166,216.55833333333334,33.46666666666667

pipeline/config.py

@@ -36,7 +36,7 @@ class QueriesConfig(Config):
 class MeasurementsConfig(Config):
     mock: bool = False
     remove_old_measurements: bool = False
-    small_dataset: bool = False
+    small_dataset: bool = True
 
 
 class QueriesDatasetConfig(Config):

pipeline/models.py

@@ -511,3 +511,20 @@ MODELS = [
         # selected=True,
     ),
 ]
+
+MISSING_MEASUREMENTS = ['Mistral-7B-Instruct-v0.2',
+ 'Vicuna v1.5 (7B)',
+ '01-ai Yi Chat (34B)',
+ 'Chronos Hermes (13B)',
+ 'LLaMA-2 Chat (7B)',
+ 'Qwen 1.5 Chat (7B)',
+ 'Snorkel Mistral PairRM DPO (7B)',
+ 'OpenHermes-2.5-Mistral (7B)',
+ 'Falcon Instruct (7B)',
+ 'RedPajama-INCITE Chat (7B)',
+ 'WizardLM v1.2 (13B)',
+ 'Upstage SOLAR Instruct v1 (11B)']
+
+for model in MODELS:
+    if model.model_display_name in MISSING_MEASUREMENTS:
+        model.selected = True