Views
No views yet
1<|im_start|>system
2system prompt<|im_end|>
3<|im_start|>user
4Hi there!<|im_end|>
5<|im_start|>assistant
6Nice to meet you!<|im_end|>
7<|im_start|>user
8Can I ask a question?<|im_end|>
9<|im_start|>assistant1{
2 "story_string": "<|im_start|>system\n{{#if system}}{{system}}\n{{/if}}{{#if wiBefore}}{{wiBefore}}\n{{/if}}{{#if description}}{{description}}\n{{/if}}{{#if personality}}{{char}}'s personality: {{personality}}\n{{/if}}{{#if scenario}}Scenario: {{scenario}}\n{{/if}}{{#if wiAfter}}{{wiAfter}}\n{{/if}}{{#if persona}}{{persona}}\n{{/if}}{{trim}}<|im_end|>\n",
3 "example_separator": "",
4 "chat_start": "",
5 "use_stop_strings": false,
6 "allow_jailbreak": false,
7 "always_force_name2": true,
8 "trim_sentences": false,
9 "include_newline": false,
10 "single_line": false,
11 "name": "Magnum ChatML"
12}1{
2 "system_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>.",
3 "input_sequence": "<|im_start|>user\n",
4 "output_sequence": "<|im_start|>assistant\n",
5 "last_output_sequence": "",
6 "system_sequence": "<|im_start|>system\n",
7 "stop_sequence": "<|im_end|>",
8 "wrap": false,
9 "macro": true,
10 "names": true,
11 "names_force_groups": true,
12 "activation_regex": "",
13 "system_sequence_prefix": "",
14 "system_sequence_suffix": "",
15 "first_output_sequence": "",
16 "skip_examples": false,
17 "output_suffix": "<|im_end|>\n",
18 "input_suffix": "<|im_end|>\n",
19 "system_suffix": "<|im_end|>\n",
20 "user_alignment_message": "",
21 "system_same_as_user": false,
22 "last_system_sequence": "",
23 "name": "Magnum ChatML"
24}1base_model: IntervitensInc/gemma-2-27b-chatml
2model_type: AutoModelForCausalLM
3tokenizer_type: AutoTokenizer
4
5hub_model_id: anthracite-org/magnum-v4-27b-r1
6hub_strategy: "all_checkpoints"
7push_dataset_to_hub:
8hf_use_auth_token: true
9
10plugins:
11 - axolotl.integrations.liger.LigerPlugin
12liger_cross_entropy: true
13#liger_rope: true
14#liger_rms_norm: true
15#liger_swiglu: true
16#liger_fused_linear_cross_entropy: true
17
18load_in_8bit: false
19load_in_4bit: false
20strict: false
21
22datasets:
23 - path: anthracite-org/c2_logs_16k_llama_v1.1
24 type: sharegpt
25 conversation: chatml
26 - path: NewEden/Claude-Instruct-5K
27 type: sharegpt
28 conversation: chatml
29 - path: anthracite-org/kalo-opus-instruct-22k-no-refusal
30 type: sharegpt
31 conversation: chatml
32 - path: Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
33 type: sharegpt
34 conversation: chatml
35 - path: lodrick-the-lafted/kalo-opus-instruct-3k-filtered
36 type: sharegpt
37 conversation: chatml
38 - path: anthracite-org/nopm_claude_writing_fixed
39 type: sharegpt
40 conversation: chatml
41 - path: Epiculous/Synthstruct-Gens-v1.1-Filtered-n-Cleaned
42 type: sharegpt
43 conversation: chatml
44 - path: anthracite-org/kalo_opus_misc_240827
45 type: sharegpt
46 conversation: chatml
47 - path: anthracite-org/kalo_misc_part2
48 type: sharegpt
49 conversation: chatml
50chat_template: chatml
51shuffle_merged_datasets: true
52default_system_message: "You are an assistant that responds to the user."
53dataset_prepared_path: /workspace/data/27-fft-data
54val_set_size: 0.0
55output_dir: /workspace/data/27b-fft-out
56
57sequence_len: 8192
58sample_packing: true
59eval_sample_packing: false
60pad_to_sequence_len: true
61
62adapter:
63lora_model_dir:
64lora_r:
65lora_alpha:
66lora_dropout:
67lora_target_linear:
68lora_fan_in_fan_out:
69
70wandb_project: 27b-nemo-config-fft
71wandb_entity:
72wandb_watch:
73wandb_name: attempt-01
74wandb_log_model:
75
76gradient_accumulation_steps: 8
77micro_batch_size: 1
78num_epochs: 4
79optimizer: paged_adamw_8bit
80lr_scheduler: cosine
81learning_rate: 0.00001
82
83train_on_inputs: false
84group_by_length: false
85bf16: auto
86fp16:
87tf32: false
88
89gradient_checkpointing: true
90early_stopping_patience:
91auto_resume_from_checkpoints: true
92local_rank:
93logging_steps: 1
94xformers_attention:
95flash_attention: true
96
97warmup_steps: 10
98evals_per_epoch:
99eval_table_size:
100eval_max_new_tokens:
101saves_per_epoch: 2
102debug:
103deepspeed: deepspeed_configs/zero3_bf16.json
104weight_decay: 0.01
105fsdp:
106fsdp_config:
107special_tokens:
108 pad_token: <pad>