Training in progress, step 10

24_10_23_config_test_3.csv

@@ -1,4 +1,4 @@
 run_number,comment,peformed_already,num_train_epochs
 1,no variations,True,2
-2,num_train_epochs set 1,False,1
+2,num_train_epochs set 1,True,1
 3,num_train_epochs set 4,False,4

24_10_23_config_test_4.csv

@@ -0,0 +1,17 @@
+run_number,comment,peformed_already,num_train_epochs,gradient_accumulation_steps,max_tokens,temperature,stop_token,max_steps
+1,no variations,True,2,4,100,0.8,False,30
+2,num_train_epochs set 1,True,1,4,100,0.8,False,30
+3,num_train_epochs set 4,True,4,4,100,0.8,False,30
+4,gradient_accumulation_steps set 2,True,2,2,100,0.8,False,30
+5,gradient_accumulation_steps set 8,False,2,8,100,0.8,False,30
+6,temperature set 0.1,False,2,4,100,0.1,False,30
+7,temperature set 0.5,False,2,4,100,0.5,False,30
+8,temperature set 0.95,False,2,4,100,0.95,False,30
+9,max_tokens set 10,False,2,4,10,0.8,False,30
+10,max_tokens set 30,False,2,4,30,0.8,False,30
+11,max_tokens set 300,False,2,4,300,0.8,False,30
+12,max_tokens set 500,False,2,4,500,0.8,False,30
+13,stop_token set True,False,2,4,100,0.8,True,30
+14,max_steps set 10,False,2,4,100,0.8,False,10
+15,max_steps set 50,False,2,4,100,0.8,False,50
+16,max_steps set 150,False,2,4,100,0.8,False,150

24_10_23_results_1st_group.csv

@@ -0,0 +1,6 @@
+run_number,items_per_minute,changed_settings,total_time_taken,rouge_scores_unnest,rouge1 low Precision,rouge1 low Recall,rouge1 low F1 Score,rouge1 mid Precision,rouge1 mid Recall,rouge1 mid F1 Score,rouge1 high Precision,rouge1 high Recall,rouge1 high F1 Score,rouge2 low Precision,rouge2 low Recall,rouge2 low F1 Score,rouge2 mid Precision,rouge2 mid Recall,rouge2 mid F1 Score,rouge2 high Precision,rouge2 high Recall,rouge2 high F1 Score
+1,1163.7459406958192,no variations,22.68536376953125,0,0.208561072443593,0.2921066803300641,0.2274536615691174,0.2497447315146519,0.3301374221663603,0.2626087517917075,0.294039439370663,0.3696421366026811,0.3008485580475517,0.0734940843610943,0.0847211653665341,0.0740754078802114,0.0991976479220887,0.1085901400630879,0.0969580865220583,0.1296589463781216,0.1378036801484575,0.125512488034304
+2,1162.808293702278,num_train_epochs set 1,22.703656435012817,0,0.1873092278292533,0.2656845537211048,0.2057998192354901,0.2275280036366529,0.3023442522081491,0.2386558568596591,0.2676587707389396,0.339965547415518,0.2730482986950854,0.0637154457398863,0.0772651769573705,0.0645895804699544,0.0860230925749903,0.1031854113096811,0.0862979830574868,0.1102734455225139,0.1284899853630821,0.1069466796729911
+3,1167.961557490705,num_train_epochs set 4,22.6034836769104,0,0.2017668631411228,0.2943660462891026,0.2275745277256124,0.2402018073982986,0.3337487272152939,0.2622230969724901,0.2800257917490945,0.3723264196525258,0.2985025000752551,0.0681875943274993,0.0902994027619567,0.0732690208470744,0.088492522537141,0.1148011192777671,0.0942046380946422,0.1095363410196173,0.143483247069912,0.1164429090551231
+4,1162.7142153040672,gradient_accumulation_steps set 2,22.705493450164795,0,0.2377524270451348,0.2987412399762022,0.2500155939457053,0.279676230316319,0.3364800986139911,0.2881491459471711,0.3237935645348746,0.3735174802097811,0.3265839526773124,0.0855126216551136,0.0949357351919399,0.0863749530085826,0.1105430616720297,0.1193682783654687,0.1089689115281058,0.1387884155416608,0.1450061706561381,0.1338769366004306
+5,0.0,0,0.0,0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0

24_10_23_results_test.csv

@@ -6,4 +6,4 @@ run_number,items_per_minute,total_time_taken,rouge_scores_unnest,rouge1 low Prec
 5,0.0,0.0,0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,,,0.0
 6,1535.3199106291877,17.19511342048645,0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,,,0.0
 7,1521.7353355093503,17.34861469268799,0,0.1675419882641928,0.2683955667174026,0.1899253049526814,0.204415224031979,0.3103849620513427,0.22453414188025,0.2480287471768989,0.3499093875751177,0.2617672658255535,0.0473700663667271,0.0678525123619495,0.0527658276302453,0.0658854770672697,0.0907404201869418,0.0723161968142757,0.0880138989246418,0.1159942948188274,0.092752111659012,,,,no variations
-8,0.0,0.0,0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,,,,0
+8,1530.9789388683143,17.243868827819824,0,0.1872570314089219,0.2547699607566171,0.2004167490068723,0.2258279932399817,0.2916713273221197,0.2337035947913197,0.2658820383391841,0.3298826790328343,0.270964750786811,0.0604505109564913,0.0735511631770191,0.0609645270385008,0.0822893827608663,0.0946933481740148,0.0806556896037959,0.106093463555273,0.1203287425439381,0.1037844035207774,,,,num_train_epochs set 1

README.md

@@ -412,6 +412,18 @@ The following `bitsandbytes` quantization config was used during training:
 - bnb_4bit_use_double_quant: True
 - bnb_4bit_compute_dtype: bfloat16
 
+The following `bitsandbytes` quantization config was used during training:
+- quant_method: bitsandbytes
+- load_in_8bit: False
+- load_in_4bit: True
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: nf4
+- bnb_4bit_use_double_quant: True
+- bnb_4bit_compute_dtype: bfloat16
+
 The following `bitsandbytes` quantization config was used during training:
 - quant_method: bitsandbytes
 - load_in_8bit: False
@@ -459,5 +471,6 @@ The following `bitsandbytes` quantization config was used during training:
 - PEFT 0.5.0
 - PEFT 0.5.0
 - PEFT 0.5.0
+- PEFT 0.5.0
 
 - PEFT 0.5.0

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0aeb59d89afea7c0f84144ec2a928e73821415ac646a42e09eda179c72c3cc17
+oid sha256:e3d6459fe040d49342ae5be1f81e84f013f524c97b97b856ada3994fb458dfdf
 size 100733709

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f71683a7af81fe9162fdfe9de6280c427004491c779477aab6335c1972c1568d
+oid sha256:6285e1a80a9f7e48077509b90c5687599369635ce95c6a9f6f78f7691632f8ac
 size 100690288

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dbfd2297e59b6e9d256eb74747a19b598047f5f2018875eb4f79b848982c63bd
-size 4283
+oid sha256:20545a7b3c481007b95896ba869ef3f39097a7d6a9b2091de4c4f169fb292497
+size 4091