See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: katuni4ka/tiny-random-falcon-40b
bf16: true
chat_template: llama3
dataset_prepared_path: null
datasets:
- data_files:
  - 9377853013503169_train_data.json
  ds_type: json
  format: custom
  path: /workspace/input_data/9377853013503169_train_data.json
  type:
    field_input: system
    field_instruction: prompt
    field_output: output
    format: '{instruction} {input}'
    no_input_format: '{instruction}'
    system_format: '{system}'
    system_prompt: ''
debug: null
deepspeed: null
early_stopping_patience: 2
eval_max_new_tokens: 128
eval_steps: 100
eval_table_size: null
flash_attention: false
fp16: null
fsdp: null
fsdp_config: null
gradient_accumulation_steps: 8
gradient_checkpointing: true
group_by_length: false
hub_model_id: Romain-XV/84f6ffaf-61c0-43d3-ab02-80efdcb5015a
hub_repo: null
hub_strategy: checkpoint
hub_token: null
learning_rate: 0.00025
load_best_model_at_end: true
load_in_4bit: false
load_in_8bit: false
local_rank: null
logging_steps: 1
lora_alpha: 64
lora_dropout: 0.1
lora_fan_in_fan_out: null
lora_model_dir: null
lora_r: 32
lora_target_linear: true
lora_target_modules:
- q_proj
- k_proj
- v_proj
- o_proj
lr_scheduler: cosine
max_grad_norm: 1.0
max_steps: 5880
micro_batch_size: 4
mlflow_experiment_name: /tmp/9377853013503169_train_data.json
model_type: AutoModelForCausalLM
num_epochs: 3
optimizer: adamw_bnb_8bit
output_dir: miner_id_24
pad_to_sequence_len: true
resume_from_checkpoint: null
s2_attention: null
sample_packing: false
save_steps: 100
sequence_len: 2048
special_tokens:
  pad_token: <|endoftext|>
strict: false
tf32: true
tokenizer_type: AutoTokenizer
train_on_inputs: false
trust_remote_code: true
val_set_size: 0.02552348671247282
wandb_entity: null
wandb_mode: online
wandb_name: 88bd9ed0-784e-40ee-bad6-32ea4fdd723a
wandb_project: Gradients-On-Demand
wandb_run: your_name
wandb_runid: 88bd9ed0-784e-40ee-bad6-32ea4fdd723a
warmup_steps: 10
weight_decay: 0.0
xformers_attention: null

84f6ffaf-61c0-43d3-ab02-80efdcb5015a

This model is a fine-tuned version of katuni4ka/tiny-random-falcon-40b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 10.5489

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.00025
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• training_steps: 5880

Training results

Training Loss	Epoch	Step	Validation Loss
88.9048	0.0002	1	11.1152
86.298	0.0168	100	10.7819
85.9568	0.0335	200	10.7421
85.6764	0.0503	300	10.7206
85.7664	0.0671	400	10.7042
85.5889	0.0838	500	10.6877
85.5986	0.1006	600	10.6747
85.3208	0.1173	700	10.6651
85.4494	0.1341	800	10.6567
85.3258	0.1509	900	10.6505
84.9632	0.1676	1000	10.6445
85.2519	0.1844	1100	10.6386
85.2217	0.2012	1200	10.6306
85.3373	0.2179	1300	10.6254
84.9467	0.2347	1400	10.6200
85.0272	0.2514	1500	10.6153
85.3044	0.2682	1600	10.6097
85.0647	0.2850	1700	10.6045
85.2542	0.3017	1800	10.5977
85.1251	0.3185	1900	10.5894
84.6675	0.3353	2000	10.5847
84.9831	0.3520	2100	10.5812
84.7303	0.3688	2200	10.5778
84.4756	0.3855	2300	10.5755
84.6714	0.4023	2400	10.5724
84.8902	0.4191	2500	10.5702
84.8641	0.4358	2600	10.5674
84.6129	0.4526	2700	10.5662
84.6396	0.4694	2800	10.5645
84.5829	0.4861	2900	10.5631
84.4782	0.5029	3000	10.5616
84.6577	0.5196	3100	10.5611
84.5671	0.5364	3200	10.5595
84.6259	0.5532	3300	10.5582
84.4862	0.5699	3400	10.5574
84.8068	0.5867	3500	10.5564
84.3837	0.6035	3600	10.5558
84.7825	0.6202	3700	10.5550
84.5206	0.6370	3800	10.5545
84.5291	0.6537	3900	10.5537
84.5388	0.6705	4000	10.5531
84.5735	0.6873	4100	10.5526
84.5167	0.7040	4200	10.5518
84.5027	0.7208	4300	10.5516
84.5048	0.7376	4400	10.5513
84.5671	0.7543	4500	10.5508
84.5944	0.7711	4600	10.5502
84.29	0.7878	4700	10.5501
84.299	0.8046	4800	10.5498
84.5276	0.8214	4900	10.5495
84.5476	0.8381	5000	10.5494
84.3845	0.8549	5100	10.5491
84.3847	0.8717	5200	10.5489
84.6318	0.8884	5300	10.5491
84.5371	0.9052	5400	10.5489
84.5082	0.9219	5500	10.5490
84.2422	0.9387	5600	10.5489
84.2555	0.9555	5700	10.5488
84.6036	0.9722	5800	10.5489

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1