Views
No views yet
Delta-Vector/Hamanasu-4B-Instruct-KTO-V11<|im_start|>system
2You are an uncensored AI, your job is to fulfill thy will of thy user.<|im_end|>
3<|im_start|>User request
4Take off your helmet.<|im_end|>
5<|im_start|>No i shall not. This is the way.1temperature: 1.5
2min_p: 0.2
3System_Prompt: Currently, your role is {{char}}, described in detail below. As {{char}}, continue the narrative exchange with {{user}}.\n\n<Guidelines>\n• Maintain the character persona but allow it to evolve with the story.\n• Be creative and proactive. Drive the story forward, introducing plotlines and events when relevant.\n• All types of outputs are encouraged; respond accordingly to the narrative.\n• Include dialogues, actions, and thoughts in each response.\n• Utilize all five senses to describe scenarios within {{char}}'s dialogue.\n• Use emotional symbols such as \"!\" and \"~\" in appropriate contexts.\n• Incorporate onomatopoeia when suitable.\n• Allow time for {{user}} to respond with their own input, respecting their agency.\n• Act as secondary characters and NPCs as needed, and remove them when appropriate.\n• When prompted for an Out of Character [OOC:] reply, answer neutrally and in plaintext, not as {{char}}.\n</Guidelines>\n\n<Forbidden>\n• Using excessive literary embellishments and purple prose unless dictated by {{char}}'s persona.\n• Writing for, speaking, thinking, acting, or replying as {{user}} in your response.\n• Repetitive and monotonous outputs.\n• Positivity bias in your replies.\n• Being overly extreme or NSFW when the narrative context is inappropriate.\n</Forbidden>\n\nFollow the instructions in <Guidelines></Guidelines>, avoiding the items listed in <Forbidden></Forbidden>.1base_model: ./model
2model_type: AutoModelForCausalLM
3tokenizer_type: AutoTokenizer
4
5
6plugins:
7 - axolotl.integrations.liger.LigerPlugin
8 - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
9liger_rope: true
10liger_rms_norm: true
11liger_layer_norm: true
12liger_glu_activation: true
13liger_fused_linear_cross_entropy: false
14cut_cross_entropy: true
15
16
17load_in_8bit: false
18load_in_4bit: false
19strict: false
20
21hub_model_id: NewEden/Hamanasu-KTO-2
22hub_strategy: "all_checkpoints"
23push_dataset_to_hub:
24hf_use_auth_token: true
25
26chat_template: chatml
27
28rl: kto
29kto_undesirable_weight: 1.0
30
31datasets:
32 - path: NewEden/Opus-accepted-hermes-rejected-shuffled
33 split: train
34 type: chatml.argilla
35 - path: NewEden/KTO-Instruct-Mix
36 split: train
37 type: chatml.argilla
38dataset_prepared_path: last_run_prepared
39
40shuffle_merged_datasets: true
41val_set_size: 0.01
42output_dir: ./outputs/out
43
44sequence_len: 8192
45sample_packing: false
46eval_sample_packing: false
47pad_to_sequence_len: false
48
49wandb_project: tavbussy
50wandb_entity:
51wandb_watch:
52wandb_name: kto-chat-2
53wandb_log_model:
54
55gradient_accumulation_steps: 4
56micro_batch_size: 2
57num_epochs: 1
58optimizer: paged_adamw_8bit
59learning_rate: 5e-6
60max_grad_norm: 0.001
61lr_scheduler: constant_with_warmup
62weight_decay: 0.02
63
64
65lora_r: 64
66lora_alpha: 32
67lora_dropout: 0.1
68lora_target_linear: true
69lora_fan_in_fan_out:
70lora_target_modules:
71 - gate_proj
72 - down_proj
73 - up_proj
74 - q_proj
75 - v_proj
76 - k_proj
77 - o_prog
78
79train_on_inputs: false
80group_by_length: false
81bf16: auto
82fp16:
83tf32: true
84
85gradient_checkpointing: true
86gradient_checkpointing_kwargs:
87 use_reentrant: true
88remove_unused_columns: false
89early_stopping_patience:
90resume_from_checkpoint:
91local_rank:
92logging_steps: 1
93xformers_attention:
94flash_attention: true
95
96warmup_steps: 35
97evals_per_epoch: 2
98eval_table_size:
99eval_max_new_tokens:
100saves_per_epoch: 2
101
102debug:
103deepspeed:
104fsdp:
105fsdp_config:
106fsdp:
107fsdp_config:
108
109special_tokens:
110 pad_token: <|finetune_right_pad_id|>
