Upload falcon-7b-custom-dpo-lora-lablebox - LoRA fine-tuned Falcon-7B

README.md

@@ -0,0 +1,91 @@
+---
+license: apache-2.0
+base_model: tiiuae/falcon-7b-instruct
+tags:
+- generated_from_trainer
+- falcon
+- lora
+- direct-preference-optimization-(custom-dpo-with-proper-loss)
+datasets:
+- HuggingFaceH4/ultrachat_200k
+- HuggingFaceH4/ultrafeedback_binarized
+metrics:
+- loss
+library_name: transformers
+model-index:
+- name: falcon-7b-custom-dpo-lora-lablebox
+  results: []
+---
+
+# falcon-7b-custom-dpo-lora-lablebox
+
+This model is a fine-tuned version of [tiiuae/falcon-7b-instruct](https://huggingface.co/tiiuae/falcon-7b-instruct) using Direct Preference Optimization (Custom DPO with Proper Loss).
+
+## Model Description
+
+- **Training Method**: Direct Preference Optimization (Custom DPO with Proper Loss)
+- **Base Model**: Falcon-7B-Instruct
+- **Parameter Count**: 6.92B (base model)
+- **LoRA Parameters**: 0.0085% trainable
+- **Hardware**: Apple Silicon Mac (128GB RAM)
+- **Framework**: PyTorch with MPS backend
+
+## Training Results
+
+- **Runtime**: 38.15 minutes
+- **Steps**: 150 optimizer steps (1200 forward passes)
+- **Loss Reduction**: 98.97%
+- **Benchmark Quality Score**: 1.00/1.00
+
+## Training Configuration
+
+### LoRA Configuration
+- Rank (r): 2
+- Alpha: 4
+- Target Modules: query_key_value
+- Dropout: 0.1
+
+### Training Parameters
+- Learning Rate: 5e-5
+- Gradient Accumulation: 8 steps
+- Mixed Precision: FP16
+- Scheduler: Cosine Annealing
+
+## Usage
+
+```python
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from peft import PeftModel
+
+# Load base model
+base_model = AutoModelForCausalLM.from_pretrained(
+    "tiiuae/falcon-7b-instruct",
+    trust_remote_code=True,
+    torch_dtype=torch.float16,
+    device_map="auto"
+)
+
+# Load LoRA adapter
+model = PeftModel.from_pretrained(base_model, "falcon-7b-custom-dpo-lora-lablebox")
+
+# Load tokenizer
+tokenizer = AutoTokenizer.from_pretrained("falcon-7b-custom-dpo-lora-lablebox")
+
+# Generate text
+prompt = "What is machine learning?"
+inputs = tokenizer(prompt, return_tensors="pt")
+outputs = model.generate(**inputs, max_length=100)
+response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+```
+
+## Training Details
+
+This model was trained as part of the Lablebox Take Home Assignment, demonstrating gradient-based training of large language models on consumer hardware.
+
+### Framework versions
+
+- Transformers 4.44.2
+- PyTorch 2.5.0.dev20240912
+- PEFT 0.13.0
+- Datasets 3.0.0
+- Tokenizers 0.19.1

adapter_config.json

@@ -0,0 +1,36 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "tiiuae/falcon-7b-instruct",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 4,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "qalora_group_size": 16,
+  "r": 2,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "query_key_value"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e7273d2e12d105705fb95348b1316044e82f7ebbe7477a7eefde52efffa375a
+size 2368664

chat_template.jinja

@@ -0,0 +1,17 @@
+{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = '' %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 %}{{ system_message.strip() }}{% endif %}{% if message['role'] == 'user' %}{{ '
+
+User: ' + message['content'].strip().replace('
+', '
+').replace('
+
+', '
+') }}{% elif message['role'] == 'assistant' %}{{ '
+
+Assistant: ' + message['content'].strip().replace('
+', '
+').replace('
+
+', '
+') }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '
+
+Assistant:' }}{% endif %}

special_tokens_map.json

@@ -0,0 +1,23 @@
+{
+  "additional_special_tokens": [
+    ">>TITLE<<",
+    ">>ABSTRACT<<",
+    ">>INTRODUCTION<<",
+    ">>SUMMARY<<",
+    ">>COMMENT<<",
+    ">>ANSWER<<",
+    ">>QUESTION<<",
+    ">>DOMAIN<<",
+    ">>PREFIX<<",
+    ">>SUFFIX<<",
+    ">>MIDDLE<<"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,124 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": ">>TITLE<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": ">>ABSTRACT<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": ">>INTRODUCTION<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": ">>SUMMARY<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": ">>COMMENT<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": ">>ANSWER<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": ">>QUESTION<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": ">>DOMAIN<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": ">>PREFIX<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "9": {
+      "content": ">>SUFFIX<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "10": {
+      "content": ">>MIDDLE<<",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "11": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    ">>TITLE<<",
+    ">>ABSTRACT<<",
+    ">>INTRODUCTION<<",
+    ">>SUMMARY<<",
+    ">>COMMENT<<",
+    ">>ANSWER<<",
+    ">>QUESTION<<",
+    ">>DOMAIN<<",
+    ">>PREFIX<<",
+    ">>SUFFIX<<",
+    ">>MIDDLE<<"
+  ],
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 128,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

training_metrics.json

@@ -0,0 +1,1219 @@
+{
+  "model_name": "tiiuae/falcon-7b-instruct",
+  "training_method": "Custom DPO (Memory Optimized)",
+  "device": "mps",
+  "final_loss": 0.007128167455084622,
+  "all_losses": [
+    0.6931445002555847,
+    0.6931498646736145,
+    0.6931474804878235,
+    0.6931471824645996,
+    0.6931498646736145,
+    0.6931471824645996,
+    0.6931471824645996,
+    0.6931483149528503,
+    0.6924462914466858,
+    0.6917656660079956,
+    0.6914777159690857,
+    0.6924195289611816,
+    0.6927425265312195,
+    0.6914411187171936,
+    0.6927124261856079,
+    0.6918513178825378,
+    0.6910051703453064,
+    0.692326545715332,
+    0.6922808289527893,
+    0.6924623250961304,
+    0.6905725002288818,
+    0.690464437007904,
+    0.6897289752960205,
+    0.6904743909835815,
+    0.6895521283149719,
+    0.6909302473068237,
+    0.6909027695655823,
+    0.6886829733848572,
+    0.6886737942695618,
+    0.6876998543739319,
+    0.6876599788665771,
+    0.6908404231071472,
+    0.6903031468391418,
+    0.6902902126312256,
+    0.6899273991584778,
+    0.6880876421928406,
+    0.6880360245704651,
+    0.6854314208030701,
+    0.6878971457481384,
+    0.6865187287330627,
+    0.6890680193901062,
+    0.6891193389892578,
+    0.6858062148094177,
+    0.6860361099243164,
+    0.6890460848808289,
+    0.6843717694282532,
+    0.682920515537262,
+    0.683175802230835,
+    0.6849693655967712,
+    0.6851096749305725,
+    0.6852080821990967,
+    0.6848581433296204,
+    0.6820118427276611,
+    0.6849530935287476,
+    0.6818994283676147,
+    0.6818982362747192,
+    0.6832471489906311,
+    0.6877012848854065,
+    0.6774458885192871,
+    0.6862993240356445,
+    0.6863565444946289,
+    0.6831058859825134,
+    0.6877350807189941,
+    0.6861698031425476,
+    0.6846379637718201,
+    0.6847556233406067,
+    0.6744025945663452,
+    0.6765581369400024,
+    0.6793200373649597,
+    0.6791293621063232,
+    0.6886783242225647,
+    0.6765052080154419,
+    0.6711891293525696,
+    0.6784105896949768,
+    0.6789045333862305,
+    0.678804874420166,
+    0.6850094199180603,
+    0.6876925826072693,
+    0.6827645897865295,
+    0.6732222437858582,
+    0.6740774512290955,
+    0.6807486414909363,
+    0.6702357530593872,
+    0.6761305928230286,
+    0.6830084919929504,
+    0.6759828925132751,
+    0.6817108392715454,
+    0.6811547875404358,
+    0.6732296943664551,
+    0.6797907948493958,
+    0.6731377243995667,
+    0.6788240075111389,
+    0.6800283193588257,
+    0.6814838647842407,
+    0.6706371903419495,
+    0.681279182434082,
+    0.6795840859413147,
+    0.6606131792068481,
+    0.6701865792274475,
+    0.6792699694633484,
+    0.6670989394187927,
+    0.6697462201118469,
+    0.6629945635795593,
+    0.6633478403091431,
+    0.66351318359375,
+    0.666574239730835,
+    0.6735833883285522,
+    0.6776118874549866,
+    0.6631135940551758,
+    0.6666763424873352,
+    0.677090048789978,
+    0.6741803288459778,
+    0.6592918038368225,
+    0.6552661061286926,
+    0.6631354689598083,
+    0.6552324295043945,
+    0.6627830862998962,
+    0.6758089661598206,
+    0.6592468619346619,
+    0.6552423238754272,
+    0.6733235716819763,
+    0.6587954163551331,
+    0.6683547496795654,
+    0.647455096244812,
+    0.6678580045700073,
+    0.6583791375160217,
+    0.6509461402893066,
+    0.6718681454658508,
+    0.671025276184082,
+    0.6652463674545288,
+    0.6504629254341125,
+    0.6693057417869568,
+    0.6692730188369751,
+    0.6462253332138062,
+    0.6497525572776794,
+    0.6423419117927551,
+    0.6501054763793945,
+    0.6654513478279114,
+    0.6773765087127686,
+    0.6501018404960632,
+    0.6614413261413574,
+    0.6409609317779541,
+    0.6615778207778931,
+    0.6615719199180603,
+    0.6626366376876831,
+    0.6455371975898743,
+    0.6574934720993042,
+    0.6626887917518616,
+    0.6576720476150513,
+    0.6625458002090454,
+    0.6458619832992554,
+    0.6354551911354065,
+    0.660053014755249,
+    0.6532688140869141,
+    0.6297677159309387,
+    0.6621207594871521,
+    0.640588104724884,
+    0.658538818359375,
+    0.6531044840812683,
+    0.6294600963592529,
+    0.6284335255622864,
+    0.6232660412788391,
+    0.6563464403152466,
+    0.6583169102668762,
+    0.658644437789917,
+    0.6585259437561035,
+    0.6486587524414062,
+    0.6583632826805115,
+    0.6510217785835266,
+    0.630347490310669,
+    0.6527251601219177,
+    0.6438708901405334,
+    0.6299727559089661,
+    0.6503925323486328,
+    0.6297138929367065,
+    0.6438133716583252,
+    0.650181770324707,
+    0.6243736147880554,
+    0.6486515402793884,
+    0.6460529565811157,
+    0.6103096008300781,
+    0.6237742900848389,
+    0.6384769082069397,
+    0.6240480542182922,
+    0.6038098335266113,
+    0.6452614068984985,
+    0.6332404613494873,
+    0.6416230201721191,
+    0.6042110919952393,
+    0.6040128469467163,
+    0.6331872344017029,
+    0.6036081910133362,
+    0.5978191494941711,
+    0.6272575855255127,
+    0.6406931281089783,
+    0.6047845482826233,
+    0.6278883814811707,
+    0.6118102669715881,
+    0.6118766665458679,
+    0.6271786093711853,
+    0.6303728818893433,
+    0.6216133832931519,
+    0.5921457409858704,
+    0.6198221445083618,
+    0.6354344487190247,
+    0.6049428582191467,
+    0.5899803042411804,
+    0.6206803917884827,
+    0.6305359601974487,
+    0.5812323093414307,
+    0.6322135329246521,
+    0.5982199311256409,
+    0.6254496574401855,
+    0.6309357285499573,
+    0.5979939103126526,
+    0.6308203339576721,
+    0.5728160738945007,
+    0.577912449836731,
+    0.5841314792633057,
+    0.579559326171875,
+    0.6086212992668152,
+    0.5909605026245117,
+    0.6278688311576843,
+    0.6253457069396973,
+    0.564061164855957,
+    0.5713092088699341,
+    0.5979952216148376,
+    0.6022064685821533,
+    0.6023690104484558,
+    0.6233775019645691,
+    0.5827073454856873,
+    0.5636537671089172,
+    0.5753647685050964,
+    0.5898715257644653,
+    0.5755640268325806,
+    0.5756675004959106,
+    0.6178175806999207,
+    0.5756856203079224,
+    0.5747778415679932,
+    0.5903384685516357,
+    0.586996853351593,
+    0.5555005669593811,
+    0.5667383670806885,
+    0.5442721247673035,
+    0.5666518211364746,
+    0.6131772398948669,
+    0.545123815536499,
+    0.6462403535842896,
+    0.607905387878418,
+    0.5788690447807312,
+    0.5776719450950623,
+    0.5526711344718933,
+    0.5344640612602234,
+    0.5575963854789734,
+    0.6075429916381836,
+    0.5787517428398132,
+    0.5439103245735168,
+    0.6017323136329651,
+    0.5481045842170715,
+    0.5938747525215149,
+    0.5648720860481262,
+    0.5480005145072937,
+    0.6017833352088928,
+    0.570075273513794,
+    0.5950820446014404,
+    0.5341002941131592,
+    0.532304048538208,
+    0.5565545558929443,
+    0.5950796008110046,
+    0.5383984446525574,
+    0.533221960067749,
+    0.6355130076408386,
+    0.5884405970573425,
+    0.5028324127197266,
+    0.5036593675613403,
+    0.6313869953155518,
+    0.630559504032135,
+    0.5042805671691895,
+    0.5815727114677429,
+    0.5232963562011719,
+    0.5385051965713501,
+    0.5132167339324951,
+    0.5420464873313904,
+    0.5391989946365356,
+    0.5424069762229919,
+    0.5750864148139954,
+    0.5136165022850037,
+    0.54123854637146,
+    0.5116230249404907,
+    0.5112494230270386,
+    0.5308729410171509,
+    0.6215913891792297,
+    0.5111666321754456,
+    0.622231662273407,
+    0.5071014165878296,
+    0.5085077285766602,
+    0.5014936327934265,
+    0.49290230870246887,
+    0.5674858689308167,
+    0.4712718725204468,
+    0.5025614500045776,
+    0.4938358664512634,
+    0.5211197733879089,
+    0.5020002126693726,
+    0.5097358822822571,
+    0.49262428283691406,
+    0.46051454544067383,
+    0.5525519847869873,
+    0.4885499179363251,
+    0.49268150329589844,
+    0.46003666520118713,
+    0.6113331317901611,
+    0.48125725984573364,
+    0.5472245812416077,
+    0.4719480276107788,
+    0.5542081594467163,
+    0.47185248136520386,
+    0.4977332651615143,
+    0.49960532784461975,
+    0.5458657145500183,
+    0.48605361580848694,
+    0.5471256375312805,
+    0.43695396184921265,
+    0.547371506690979,
+    0.5411002039909363,
+    0.5374128818511963,
+    0.48689737915992737,
+    0.5411355495452881,
+    0.4556392729282379,
+    0.4745101034641266,
+    0.5939860939979553,
+    0.4611460864543915,
+    0.4621794819831848,
+    0.5286901593208313,
+    0.4262191355228424,
+    0.4488903880119324,
+    0.43746116757392883,
+    0.5253694653511047,
+    0.4135725498199463,
+    0.4603022634983063,
+    0.5876001119613647,
+    0.4704902470111847,
+    0.43742114305496216,
+    0.4147413969039917,
+    0.4418703019618988,
+    0.5819816589355469,
+    0.5160590410232544,
+    0.4605996012687683,
+    0.40328705310821533,
+    0.44819375872612,
+    0.582499086856842,
+    0.40257078409194946,
+    0.5744220018386841,
+    0.5161323547363281,
+    0.42198461294174194,
+    0.4507005512714386,
+    0.4505624771118164,
+    0.389405220746994,
+    0.42344269156455994,
+    0.4131486713886261,
+    0.4112369120121002,
+    0.5080081820487976,
+    0.41043415665626526,
+    0.42155003547668457,
+    0.4015685021877289,
+    0.49989670515060425,
+    0.3771463632583618,
+    0.49973490834236145,
+    0.3651106655597687,
+    0.388476699590683,
+    0.4131292402744293,
+    0.38985753059387207,
+    0.41296517848968506,
+    0.4875871241092682,
+    0.413534551858902,
+    0.4903828203678131,
+    0.3772355318069458,
+    0.39644503593444824,
+    0.490755558013916,
+    0.491123229265213,
+    0.4911811351776123,
+    0.4031429588794708,
+    0.37661072611808777,
+    0.4787423312664032,
+    0.48106497526168823,
+    0.39308962225914,
+    0.4690783619880676,
+    0.33858534693717957,
+    0.39356479048728943,
+    0.3379952013492584,
+    0.4061875641345978,
+    0.39284342527389526,
+    0.37173643708229065,
+    0.3626065254211426,
+    0.3514269292354584,
+    0.36128658056259155,
+    0.3965463936328888,
+    0.47016477584838867,
+    0.383629709482193,
+    0.3255033493041992,
+    0.35901546478271484,
+    0.45945486426353455,
+    0.3588974177837372,
+    0.3722594976425171,
+    0.34977397322654724,
+    0.3586667478084564,
+    0.3726046681404114,
+    0.33926698565483093,
+    0.34642764925956726,
+    0.33725613355636597,
+    0.32578474283218384,
+    0.3012453615665436,
+    0.3619304299354553,
+    0.3619183897972107,
+    0.33702749013900757,
+    0.44183972477912903,
+    0.33244702219963074,
+    0.3148934841156006,
+    0.36402735114097595,
+    0.3240016996860504,
+    0.3232402503490448,
+    0.43226319551467896,
+    0.4344284236431122,
+    0.3638581335544586,
+    0.3404121994972229,
+    0.42569249868392944,
+    0.42337852716445923,
+    0.302324116230011,
+    0.3398860692977905,
+    0.3185364305973053,
+    0.34056708216667175,
+    0.3027110695838928,
+    0.41399940848350525,
+    0.29627907276153564,
+    0.41602903604507446,
+    0.32866451144218445,
+    0.41556599736213684,
+    0.2914108335971832,
+    0.3284965455532074,
+    0.29041633009910583,
+    0.49763035774230957,
+    0.29284006357192993,
+    0.4037092626094818,
+    0.29212823510169983,
+    0.27949556708335876,
+    0.28409966826438904,
+    0.4046476185321808,
+    0.29120299220085144,
+    0.3919215798377991,
+    0.30811893939971924,
+    0.27066531777381897,
+    0.49170199036598206,
+    0.3195757269859314,
+    0.26983997225761414,
+    0.2669019103050232,
+    0.27848219871520996,
+    0.2547789514064789,
+    0.2974504828453064,
+    0.26666125655174255,
+    0.23456233739852905,
+    0.2559671998023987,
+    0.38447028398513794,
+    0.38289400935173035,
+    0.3107762932777405,
+    0.2988559305667877,
+    0.24448160827159882,
+    0.3724191188812256,
+    0.2874979078769684,
+    0.25385788083076477,
+    0.4768151640892029,
+    0.47619226574897766,
+    0.28776121139526367,
+    0.36527010798454285,
+    0.24211189150810242,
+    0.2317088097333908,
+    0.27747198939323425,
+    0.2773359417915344,
+    0.2762437164783478,
+    0.27781257033348083,
+    0.2870321571826935,
+    0.22294077277183533,
+    0.205506831407547,
+    0.2658420205116272,
+    0.23179513216018677,
+    0.2660740911960602,
+    0.2657136023044586,
+    0.26531633734703064,
+    0.3500315845012665,
+    0.2546113133430481,
+    0.34498539566993713,
+    0.34042835235595703,
+    0.22066892683506012,
+    0.3472563624382019,
+    0.2131386250257492,
+    0.20853550732135773,
+    0.21361389756202698,
+    0.44278255105018616,
+    0.32870081067085266,
+    0.32989436388015747,
+    0.24386601150035858,
+    0.24373729526996613,
+    0.20331601798534393,
+    0.19704888761043549,
+    0.24429944157600403,
+    0.18628498911857605,
+    0.23279482126235962,
+    0.17954613268375397,
+    0.31857889890670776,
+    0.3170939087867737,
+    0.24498921632766724,
+    0.3257792294025421,
+    0.18555884063243866,
+    0.4273545742034912,
+    0.3161287307739258,
+    0.2216482013463974,
+    0.18559527397155762,
+    0.30575141310691833,
+    0.2216119021177292,
+    0.42628270387649536,
+    0.22195522487163544,
+    0.30648308992385864,
+    0.16562257707118988,
+    0.22686660289764404,
+    0.17801722884178162,
+    0.183041051030159,
+    0.29550305008888245,
+    0.21192996203899384,
+    0.21171073615550995,
+    0.1536053717136383,
+    0.16992346942424774,
+    0.1533624231815338,
+    0.2846032977104187,
+    0.15464404225349426,
+    0.289499968290329,
+    0.15336313843727112,
+    0.1748756319284439,
+    0.14590901136398315,
+    0.1667328178882599,
+    0.2733495831489563,
+    0.28692227602005005,
+    0.19089758396148682,
+    0.16649959981441498,
+    0.14543405175209045,
+    0.19044874608516693,
+    0.18160457909107208,
+    0.18118010461330414,
+    0.181069016456604,
+    0.3883045017719269,
+    0.261026531457901,
+    0.18171244859695435,
+    0.19860412180423737,
+    0.15499143302440643,
+    0.2673388123512268,
+    0.15064392983913422,
+    0.3796258866786957,
+    0.1885426938533783,
+    0.1719590723514557,
+    0.26763731241226196,
+    0.17158840596675873,
+    0.2675325870513916,
+    0.12149936705827713,
+    0.1628296971321106,
+    0.257920503616333,
+    0.14409102499485016,
+    0.16261634230613708,
+    0.14411258697509766,
+    0.11936276406049728,
+    0.1442900002002716,
+    0.15381605923175812,
+    0.1538545787334442,
+    0.17111986875534058,
+    0.13673533499240875,
+    0.15364094078540802,
+    0.13648462295532227,
+    0.15382085740566254,
+    0.11248353868722916,
+    0.1284705549478531,
+    0.10767433792352676,
+    0.1623470038175583,
+    0.1615772843360901,
+    0.21944770216941833,
+    0.1622009575366974,
+    0.12895824015140533,
+    0.13029201328754425,
+    0.1529310941696167,
+    0.2316240668296814,
+    0.22977925837039948,
+    0.12518227100372314,
+    0.10138191282749176,
+    0.20921847224235535,
+    0.10204995423555374,
+    0.153370201587677,
+    0.11576832085847855,
+    0.332719624042511,
+    0.14384855329990387,
+    0.22274072468280792,
+    0.09217941761016846,
+    0.12970037758350372,
+    0.22235463559627533,
+    0.092031329870224,
+    0.21452496945858002,
+    0.08616142719984055,
+    0.19010883569717407,
+    0.08598838746547699,
+    0.13536641001701355,
+    0.11477453261613846,
+    0.12270762771368027,
+    0.19011740386486053,
+    0.08391938358545303,
+    0.10413017123937607,
+    0.11583764851093292,
+    0.12667758762836456,
+    0.11639966070652008,
+    0.18030133843421936,
+    0.0807802826166153,
+    0.2086164951324463,
+    0.09767240285873413,
+    0.19968107342720032,
+    0.17134496569633484,
+    0.11898531019687653,
+    0.09840808063745499,
+    0.10531356185674667,
+    0.0979539006948471,
+    0.20009057223796844,
+    0.10196167975664139,
+    0.103786900639534,
+    0.10384460538625717,
+    0.07066081464290619,
+    0.11088051646947861,
+    0.194127157330513,
+    0.07415738701820374,
+    0.10151051729917526,
+    0.09781283885240555,
+    0.09813503175973892,
+    0.28520989418029785,
+    0.15282398462295532,
+    0.09743776172399521,
+    0.28450801968574524,
+    0.183410182595253,
+    0.1844703108072281,
+    0.09261145442724228,
+    0.09259688854217529,
+    0.06175907701253891,
+    0.09248984605073929,
+    0.06517014652490616,
+    0.144436314702034,
+    0.06500717997550964,
+    0.0985884889960289,
+    0.06170068681240082,
+    0.08891043066978455,
+    0.060722120106220245,
+    0.17274008691310883,
+    0.08725526928901672,
+    0.17030903697013855,
+    0.08745495229959488,
+    0.0779389962553978,
+    0.08279872685670853,
+    0.16612842679023743,
+    0.0848052129149437,
+    0.08498108386993408,
+    0.08733031898736954,
+    0.0538545660674572,
+    0.05413297191262245,
+    0.07482273876667023,
+    0.08140134811401367,
+    0.07808542251586914,
+    0.15740883350372314,
+    0.05060839280486107,
+    0.05062108486890793,
+    0.07066836953163147,
+    0.15739069879055023,
+    0.07058772444725037,
+    0.11598871648311615,
+    0.07767332345247269,
+    0.07423298805952072,
+    0.07410160452127457,
+    0.0661250650882721,
+    0.15170904994010925,
+    0.07419092208147049,
+    0.1536543369293213,
+    0.07391837984323502,
+    0.07472262531518936,
+    0.0470963679254055,
+    0.14525943994522095,
+    0.07434919476509094,
+    0.07462549209594727,
+    0.06316735595464706,
+    0.07398713380098343,
+    0.10506890714168549,
+    0.07109792530536652,
+    0.13960401713848114,
+    0.06049637123942375,
+    0.1400953233242035,
+    0.07119550555944443,
+    0.14230406284332275,
+    0.06005876511335373,
+    0.13698899745941162,
+    0.0631227046251297,
+    0.06297764927148819,
+    0.03925151005387306,
+    0.06299925595521927,
+    0.06325296312570572,
+    0.13649646937847137,
+    0.0629970133304596,
+    0.05378938838839531,
+    0.06450459361076355,
+    0.208353653550148,
+    0.09393519908189774,
+    0.1300305724143982,
+    0.20770245790481567,
+    0.06392695009708405,
+    0.038912978023290634,
+    0.03449998050928116,
+    0.06094411015510559,
+    0.19998419284820557,
+    0.05629725754261017,
+    0.08995649963617325,
+    0.05070539563894272,
+    0.05099249631166458,
+    0.05631033331155777,
+    0.05361461266875267,
+    0.05806032568216324,
+    0.04804699867963791,
+    0.05332792550325394,
+    0.05387614294886589,
+    0.05323614925146103,
+    0.05815042555332184,
+    0.053521960973739624,
+    0.04561753198504448,
+    0.08139926195144653,
+    0.054878465831279755,
+    0.0551244392991066,
+    0.0456833578646183,
+    0.045276131480932236,
+    0.03047638013958931,
+    0.1159338429570198,
+    0.11017680913209915,
+    0.02905959077179432,
+    0.04293733090162277,
+    0.11113686859607697,
+    0.02888599969446659,
+    0.031032411381602287,
+    0.043122775852680206,
+    0.10938327759504318,
+    0.10555075854063034,
+    0.040388934314250946,
+    0.027218958362936974,
+    0.027276845648884773,
+    0.05162633955478668,
+    0.07320689409971237,
+    0.104668527841568,
+    0.10550194978713989,
+    0.10326311737298965,
+    0.04261729493737221,
+    0.0477943941950798,
+    0.07015129178762436,
+    0.03817280754446983,
+    0.07003726065158844,
+    0.028019791468977928,
+    0.049845870584249496,
+    0.04106797277927399,
+    0.06692957878112793,
+    0.15731248259544373,
+    0.06649645417928696,
+    0.04777837172150612,
+    0.03588022291660309,
+    0.10032788664102554,
+    0.048243198543787,
+    0.034380316734313965,
+    0.03860827907919884,
+    0.022958889603614807,
+    0.022758832201361656,
+    0.03870966657996178,
+    0.03384919464588165,
+    0.03843679651618004,
+    0.04562322422862053,
+    0.03202221170067787,
+    0.03203202411532402,
+    0.021567918360233307,
+    0.032222963869571686,
+    0.06020233407616615,
+    0.14640145003795624,
+    0.08551933616399765,
+    0.0856010839343071,
+    0.03489112854003906,
+    0.020358998328447342,
+    0.08126595616340637,
+    0.042565543204545975,
+    0.09101268649101257,
+    0.023345721885561943,
+    0.03472856059670448,
+    0.05675363540649414,
+    0.05414426326751709,
+    0.05414753779768944,
+    0.01939511112868786,
+    0.08808095753192902,
+    0.02856328710913658,
+    0.03307066485285759,
+    0.08813313394784927,
+    0.038151804357767105,
+    0.1282176524400711,
+    0.03115582838654518,
+    0.01850273460149765,
+    0.03968880698084831,
+    0.1289437711238861,
+    0.07305321097373962,
+    0.07344783842563629,
+    0.031223079189658165,
+    0.03465590626001358,
+    0.034709103405475616,
+    0.020741933956742287,
+    0.02549893409013748,
+    0.01725928485393524,
+    0.08208199590444565,
+    0.017377501353621483,
+    0.025658823549747467,
+    0.046369172632694244,
+    0.02428009919822216,
+    0.019997181370854378,
+    0.02836006134748459,
+    0.03314285725355148,
+    0.028280219063162804,
+    0.024202723056077957,
+    0.03662899136543274,
+    0.03557897359132767,
+    0.026715703308582306,
+    0.04420220106840134,
+    0.035717934370040894,
+    0.035446204245090485,
+    0.11303487420082092,
+    0.02687249891459942,
+    0.03556240722537041,
+    0.07435694336891174,
+    0.018380669876933098,
+    0.025674276053905487,
+    0.03018934465944767,
+    0.025509392842650414,
+    0.07492759823799133,
+    0.021768679842352867,
+    0.025608163326978683,
+    0.10356583446264267,
+    0.024305695667862892,
+    0.07274787873029709,
+    0.02021040767431259,
+    0.01732548326253891,
+    0.014172733761370182,
+    0.10311657935380936,
+    0.04015694931149483,
+    0.028021298348903656,
+    0.02302529476583004,
+    0.0693942978978157,
+    0.02783268690109253,
+    0.019521726295351982,
+    0.016449080780148506,
+    0.03829626366496086,
+    0.02777726575732231,
+    0.015750879421830177,
+    0.022011732682585716,
+    0.01844913512468338,
+    0.030725881457328796,
+    0.05375843495130539,
+    0.016024962067604065,
+    0.0221174955368042,
+    0.012806553393602371,
+    0.02082611620426178,
+    0.06471490859985352,
+    0.06495456397533417,
+    0.05105230584740639,
+    0.029913099482655525,
+    0.01761529967188835,
+    0.02101733908057213,
+    0.020816192030906677,
+    0.0118758799508214,
+    0.024353278800845146,
+    0.03329184278845787,
+    0.028668126091361046,
+    0.08612500876188278,
+    0.06303495913743973,
+    0.01996738463640213,
+    0.0198186207562685,
+    0.011274555698037148,
+    0.03189116716384888,
+    0.060677576810121536,
+    0.011297069489955902,
+    0.013953866437077522,
+    0.018967796117067337,
+    0.013768225908279419,
+    0.0608196035027504,
+    0.01799273118376732,
+    0.018005844205617905,
+    0.02705247327685356,
+    0.05832856893539429,
+    0.018186602741479874,
+    0.045446332544088364,
+    0.01523747481405735,
+    0.01527586579322815,
+    0.02588466927409172,
+    0.014593173749744892,
+    0.02603543922305107,
+    0.07579998672008514,
+    0.02130398526787758,
+    0.07544110715389252,
+    0.05617983639240265,
+    0.017088552936911583,
+    0.025242196395993233,
+    0.05434073880314827,
+    0.011736392974853516,
+    0.016324538737535477,
+    0.013875570148229599,
+    0.00995834730565548,
+    0.02501341514289379,
+    0.013839947059750557,
+    0.024364331737160683,
+    0.019488992169499397,
+    0.052454929798841476,
+    0.019539261236786842,
+    0.015594560652971268,
+    0.015654180198907852,
+    0.009525090456008911,
+    0.040793001651763916,
+    0.01281232014298439,
+    0.023437952622771263,
+    0.012722281739115715,
+    0.00927579402923584,
+    0.010758965276181698,
+    0.023302050307393074,
+    0.010716980323195457,
+    0.009196422062814236,
+    0.012283862568438053,
+    0.022570855915546417,
+    0.012104624882340431,
+    0.010285703465342522,
+    0.063756063580513,
+    0.048849720507860184,
+    0.02277025207877159,
+    0.012194129638373852,
+    0.013649464584887028,
+    0.061110541224479675,
+    0.009697116911411285,
+    0.013558678328990936,
+    0.03617057576775551,
+    0.011763019487261772,
+    0.02204589918255806,
+    0.013672629371285439,
+    0.008239207789301872,
+    0.020937608554959297,
+    0.023888971656560898,
+    0.01295860018581152,
+    0.0082494942471385,
+    0.009257960133254528,
+    0.02342071942985058,
+    0.045510124415159225,
+    0.022909024730324745,
+    0.016102973371744156,
+    0.008905809372663498,
+    0.01584017649292946,
+    0.044379424303770065,
+    0.007914611138403416,
+    0.012482728809118271,
+    0.03377128764986992,
+    0.04281456023454666,
+    0.007688809186220169,
+    0.01945801079273224,
+    0.011936899274587631,
+    0.053659241646528244,
+    0.007626701612025499,
+    0.007673193700611591,
+    0.008628113195300102,
+    0.010137615725398064,
+    0.02155928499996662,
+    0.014737549237906933,
+    0.007409236393868923,
+    0.010095249861478806,
+    0.008118250407278538,
+    0.014707712456583977,
+    0.014681986533105373,
+    0.010801891796290874,
+    0.007066851016134024,
+    0.007132543716579676,
+    0.018458731472492218,
+    0.009614826180040836,
+    0.007030273787677288,
+    0.030689697712659836,
+    0.00971116591244936,
+    0.007582573220133781,
+    0.007560212630778551,
+    0.009404760785400867,
+    0.018014390021562576,
+    0.018174076452851295,
+    0.010439195670187473,
+    0.04770302027463913,
+    0.01982855424284935,
+    0.0384221151471138,
+    0.017519133165478706,
+    0.045681536197662354,
+    0.006694271694868803,
+    0.010092182084918022,
+    0.006609838921576738,
+    0.01750226505100727,
+    0.006641930900514126,
+    0.018852457404136658,
+    0.008757871575653553,
+    0.008739910088479519,
+    0.006435384973883629,
+    0.006513910833746195,
+    0.006421882193535566,
+    0.006908577401190996,
+    0.016923654824495316,
+    0.012021001428365707,
+    0.018463412299752235,
+    0.011996736750006676,
+    0.009289612993597984,
+    0.008412993513047695,
+    0.016532570123672485,
+    0.006137215532362461,
+    0.04224630445241928,
+    0.02647797204554081,
+    0.008114348165690899,
+    0.00805013906210661,
+    0.03560060262680054,
+    0.008099686354398727,
+    0.016109425574541092,
+    0.008110328577458858,
+    0.0357891321182251,
+    0.011022393591701984,
+    0.005766778718680143,
+    0.011249330826103687,
+    0.0058521125465631485,
+    0.039680443704128265,
+    0.007769601885229349,
+    0.008561336435377598,
+    0.006277486216276884,
+    0.010645860806107521,
+    0.005558155011385679,
+    0.005592889152467251,
+    0.007631315384060144,
+    0.007545660249888897,
+    0.00557309202849865,
+    0.01510127354413271,
+    0.007509337738156319,
+    0.024525102227926254,
+    0.014727800153195858,
+    0.032998692244291306,
+    0.00798970926553011,
+    0.02444111369550228,
+    0.0059145670384168625,
+    0.01467352919280529,
+    0.007273266091942787,
+    0.007598663680255413,
+    0.03250468894839287,
+    0.005221894942224026,
+    0.007062233984470367,
+    0.014368993230164051,
+    0.005186317954212427,
+    0.007637230679392815,
+    0.0071727861650288105,
+    0.00553883146494627,
+    0.015408169478178024,
+    0.0316757969558239,
+    0.006800838280469179,
+    0.03392940014600754,
+    0.005596089642494917,
+    0.006855893414467573,
+    0.013935409486293793,
+    0.013600778765976429,
+    0.009333429858088493,
+    0.0066059306263923645,
+    0.013660752214491367,
+    0.022320829331874847,
+    0.006648325826972723,
+    0.005374981090426445,
+    0.03280490264296532,
+    0.006788879632949829,
+    0.006889635231345892,
+    0.006878743413835764,
+    0.004766055848449469,
+    0.004800579976290464,
+    0.006821677088737488,
+    0.021795623004436493,
+    0.006406958214938641,
+    0.008698076009750366,
+    0.006594325415790081,
+    0.0064928289502859116,
+    0.006461797747761011,
+    0.029628215357661247,
+    0.00619669072329998,
+    0.030427930876612663,
+    0.029721025377511978,
+    0.00491992337629199,
+    0.02072185091674328,
+    0.013841710053384304,
+    0.014048143289983273,
+    0.006337308324873447,
+    0.006019795313477516,
+    0.012615637853741646,
+    0.012568551115691662,
+    0.012295755557715893,
+    0.005756585858762264,
+    0.028263067826628685,
+    0.0048377132043242455,
+    0.008061964064836502,
+    0.02834777720272541,
+    0.006076196674257517,
+    0.004772106651216745,
+    0.027578983455896378,
+    0.027783064171671867,
+    0.027765672653913498,
+    0.027766946703195572,
+    0.0076950788497924805,
+    0.0058954874984920025,
+    0.005574633367359638,
+    0.027760570868849754,
+    0.004509873688220978,
+    0.01162351667881012,
+    0.00566982151940465,
+    0.005404504481703043,
+    0.012854570522904396,
+    0.004122450482100248,
+    0.00753027992323041,
+    0.011642840690910816,
+    0.01138204988092184,
+    0.005541558377444744,
+    0.005255692172795534,
+    0.005272886715829372,
+    0.005462837405502796,
+    0.005315694026648998,
+    0.005482162348926067,
+    0.005250118672847748,
+    0.00392037071287632,
+    0.01218883041292429,
+    0.004222762770950794,
+    0.012111103162169456,
+    0.005280120298266411,
+    0.012218507938086987,
+    0.011025813408195972,
+    0.005098673980683088,
+    0.006903013680130243,
+    0.004197834059596062,
+    0.004231665749102831,
+    0.011809554882347584,
+    0.010888329707086086,
+    0.005138760898262262,
+    0.004948985762894154,
+    0.010789037682116032,
+    0.01057190541177988,
+    0.011509330943226814,
+    0.0050407941453158855,
+    0.006610786076635122,
+    0.024810977280139923,
+    0.0050112600438296795,
+    0.024488458409905434,
+    0.004096213262528181,
+    0.022485073655843735,
+    0.004858473781496286,
+    0.023943325504660606,
+    0.0036052961368113756,
+    0.004811969585716724,
+    0.024021420627832413,
+    0.00482252798974514,
+    0.022659895941615105,
+    0.023460427299141884,
+    0.010769696906208992,
+    0.004542389884591103,
+    0.016600457951426506,
+    0.003502308391034603,
+    0.003516444470733404,
+    0.004586771130561829,
+    0.004557935055345297,
+    0.004601841326802969,
+    0.006131173111498356,
+    0.003474866971373558,
+    0.009830987080931664,
+    0.022840164601802826,
+    0.02292405441403389,
+    0.01061861403286457,
+    0.021145032718777657,
+    0.005910301115363836,
+    0.0036365347914397717,
+    0.009599831886589527,
+    0.0033308761194348335,
+    0.004276416730135679,
+    0.015887577086687088,
+    0.0036367725115269423,
+    0.009621202014386654,
+    0.015479302033782005,
+    0.00432888139039278,
+    0.004331374075263739,
+    0.004275348503142595,
+    0.0042358203791081905,
+    0.015305684879422188,
+    0.0034895974677056074,
+    0.0032998654060065746,
+    0.0031985098030418158,
+    0.015040447004139423,
+    0.004169106017798185,
+    0.015210000798106194,
+    0.015159747563302517,
+    0.004115802235901356,
+    0.003438158193603158,
+    0.004160440061241388
+  ],
+  "train_runtime": 2288.977719068527,
+  "train_runtime_minutes": 38.14962865114212,
+  "train_steps": 1200,
+  "learning_rate": 5e-05,
+  "beta": 0.1,
+  "lora_r": 2,
+  "lora_alpha": 4,
+  "max_length": 128,
+  "status": "SUCCESS",
+  "proper_dpo_loss": true,
+  "output_dir": "models/falcon_dpo_custom_v2"
+}