Views
No views yet
0.5.01#base_model: unsloth/Llama-3.2-3B
2#base_model: anthracite-core/llama3.2-3b-chatml-v2
3base_model: NewEden/L3.2-Erebus-Base
4model_type: AutoModelForCausalLM
5tokenizer_type: AutoTokenizer
6
7load_in_8bit: false
8load_in_4bit: false
9strict: false
10
11datasets:
12 - path: NewEden/Claude-Instruct-5K
13 type: chat_template
14 chat_template: chatml
15 roles_to_train: ["gpt"]
16 field_messages: conversations
17 message_field_role: from
18 message_field_content: value
19 train_on_eos: turn
20 - path: NewEden/Claude-Instruct-2.7K
21 type: chat_template
22 chat_template: chatml
23 roles_to_train: ["gpt"]
24 field_messages: conversations
25 message_field_role: from
26 message_field_content: value
27 train_on_eos: turn
28 - path: anthracite-org/kalo-opus-instruct-22k-no-refusal
29 type: chat_template
30 chat_template: chatml
31 roles_to_train: ["gpt"]
32 field_messages: conversations
33 message_field_role: from
34 message_field_content: value
35 train_on_eos: turn
36
37chat_template: chatml
38shuffle_merged_datasets: true
39default_system_message: "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>."
40dataset_prepared_path: 4b-erebus-lora
41val_set_size: 0.0
42output_dir: 4b-erebus-rslora-out
43
44plugins:
45 - axolotl.integrations.liger.LigerPlugin
46liger_rope: true
47liger_rms_norm: true
48liger_swiglu: true
49liger_cross_entropy: false
50liger_fused_linear_cross_entropy: true
51
52sequence_len: 8192
53sample_packing: true
54pad_to_sequence_len: true
55
56wandb_project: 4b-erebus
57wandb_entity:
58wandb_watch:
59wandb_name: base-attempt-01
60wandb_log_model:
61
62gradient_accumulation_steps: 16
63micro_batch_size: 1
64num_epochs: 2
65optimizer: paged_adamw_8bit
66lr_scheduler: cosine
67learning_rate: 0.0001
68
69train_on_inputs: false
70group_by_length: false
71bf16: auto
72fp16:
73tf32: false
74
75gradient_checkpointing: unsloth
76early_stopping_patience:
77resume_from_checkpoint:
78local_rank:
79logging_steps: 1
80xformers_attention:
81flash_attention: true
82
83warmup_steps: 40
84evals_per_epoch:
85eval_table_size:
86eval_max_new_tokens:
87saves_per_epoch: 2
88debug:
89#deepspeed: /workspace/axolotl/deepspeed_configs/zero3_bf16_cpuoffload_params.json
90weight_decay: 0.01
91fsdp:
92fsdp_config:
93special_tokens:
94 pad_token: <|finetune_right_pad_id|>
95 eos_token: <|eot_id|>
96