You can deploy this model and start earning money today!

RichardErkhov/
anthracite-org_-_magnum-v2-4b-gguf

Name: anthracite-org_-_magnum-v2-4b-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

magnum-v2-4b - GGUF

Model creator: https://huggingface.co/anthracite-org/
Original model: https://huggingface.co/anthracite-org/magnum-v2-4b/

Name	Quant method	Size
magnum-v2-4b.Q2_K.gguf	Q2_K	1.71GB
magnum-v2-4b.IQ3_XS.gguf	IQ3_XS	1.89GB
magnum-v2-4b.IQ3_S.gguf	IQ3_S	1.97GB
magnum-v2-4b.Q3_K_S.gguf	Q3_K_S	1.96GB
magnum-v2-4b.IQ3_M.gguf	IQ3_M	2.03GB
magnum-v2-4b.Q3_K.gguf	Q3_K	2.14GB
magnum-v2-4b.Q3_K_M.gguf	Q3_K_M	2.14GB
magnum-v2-4b.Q3_K_L.gguf	Q3_K_L	2.3GB
magnum-v2-4b.IQ4_XS.gguf	IQ4_XS	2.38GB
magnum-v2-4b.Q4_0.gguf	Q4_0	2.47GB
magnum-v2-4b.IQ4_NL.gguf	IQ4_NL	2.49GB
magnum-v2-4b.Q4_K_S.gguf	Q4_K_S	2.48GB
magnum-v2-4b.Q4_K.gguf	Q4_K	2.59GB
magnum-v2-4b.Q4_K_M.gguf	Q4_K_M	2.59GB
magnum-v2-4b.Q4_1.gguf	Q4_1	2.71GB
magnum-v2-4b.Q5_0.gguf	Q5_0	2.95GB
magnum-v2-4b.Q5_K_S.gguf	Q5_K_S	2.95GB
magnum-v2-4b.Q5_K.gguf	Q5_K	3.01GB
magnum-v2-4b.Q5_K_M.gguf	Q5_K_M	3.01GB
magnum-v2-4b.Q5_1.gguf	Q5_1	3.19GB
magnum-v2-4b.Q6_K.gguf	Q6_K	3.46GB
magnum-v2-4b.Q8_0.gguf	Q8_0	4.47GB

Original model description:

license: apache-2.0 language:

en pipeline_tag: text-generation base_model: nvidia/Llama-3.1-Minitron-4B-Width-Base tags:
chat

This is the eighth in a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet and Opus. This model is fine-tuned on top of IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml.

Prompting

Model has been Instruct tuned with the ChatML formatting. A typical input would look like this:

"""<|im_start|>system
system prompt<|im_end|>
<|im_start|>user
Hi there!<|im_end|>
<|im_start|>assistant
Nice to meet you!<|im_end|>
<|im_start|>user
Can I ask a question?<|im_end|>
<|im_start|>assistant
"""

axolotl config

See axolotl config

axolotl version: 0.4.1

yaml

base_model: IntervitensInc/Llama-3.1-Minitron-4B-Width-Base-chatml
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: anthracite-org/Gryphe-3.5-16k-Subset
    type: sharegpt
    conversation: chatml
  - path: Epiculous/Synthstruct-Gens-v1-Filtered-n-Cleaned
    type: sharegpt
    conversation: chatml
  - path: anthracite-org/Stheno-Data-Filtered
    type: sharegpt
    conversation: chatml
  - path: Epiculous/SynthRP-Gens-v1-Filtered-n-Cleaned
    type: sharegpt
    conversation: chatml
  - path: lodrick-the-lafted/NopmWritingStruct
    type: sharegpt
    conversation: chatml
  - path: anthracite-org/kalo-opus-instruct-22k-no-refusal
    type: sharegpt
    conversation: chatml

chat_template: chatml

val_set_size: 0.01
output_dir: ./outputs/out

adapter:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:

sequence_len: 16384
# sequence_len: 32768
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 32
micro_batch_size: 1
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.00002
weight_decay: 0.05

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: true

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_ratio: 0.1
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1

debug:
deepspeed:
fsdp:
fsdp_config:

special_tokens:
  pad_token: <|finetune_right_pad_id|>

Credits

This model has been a team effort, and the credits goes to all members of Anthracite.

Training

The training was done for 2 epochs. We used 2 x RTX 6000s GPUs graciously provided by Kubernetes_Bad for the full-parameter fine-tuning of the model.

Safety

...

RichardErkhov/anthracite-org_-_magnum-v2-4b-gguf