π Model Description
license: other base_model: cognitivecomputations/dolphin-2.9-llama3-8b tags:
- generatedfromtrainer
- name: out
- cognitivecomputations/Dolphin-2.9
- teknium/OpenHermes-2.5
- m-a-p/CodeFeedback-Filtered-Instruction
- cognitivecomputations/dolphin-coder
- cognitivecomputations/samantha-data
- HuggingFaceH4/ultrachat_200k
- microsoft/orca-math-word-problems-200k
- abacusai/SystemChat-1.1
- Locutusque/function-calling-chatml
- internlm/Agent-FLAN
Dolphin 2.9 Llama 3 8b- GGUF π¬
- This is GGUF quantized version of cognitivecomputations/dolphin-2.9-llama3-8b created using llama.cpp
- Curated and trained by Eric Hartford, Lucas Atkins, and Fernando Fernandes, and Cognitive Computations
Model Description
My appreciation for the sponsors of Dolphin 2.9:- Crusoe Cloud - provided excellent on-demand 10xL40S node
This model is based on Llama-3-8b, and is governed by META LLAMA 3 COMMUNITY LICENSE AGREEMENT
The base model has 8k context, and the full-weight fine-tuning was with 4k sequence length.
It took 2.5 days on 8x L40S provided by Crusoe Cloud
This model was trained FFT on all parameters, using ChatML prompt template format.
example:
<|im_start|>system
You are Dolphin, a helpful AI assistant.<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistantDolphin-2.9 has a variety of instruction, conversational, and coding skills. It also has initial agentic abilities and supports function calling.
Dolphin is uncensored. I have filtered the dataset to remove alignment and bias. This makes the model more compliant. You are advised to implement your own alignment layer before exposing the model as a service. It will be highly compliant with any requests, even unethical ones. Please read my blog post about uncensored models. https://erichartford.com/uncensored-models You are responsible for any content you create using this model. Enjoy responsibly.
Dolphin is licensed according to Meta's Llama license. I grant permission for any use, including commercial, that falls within accordance with Meta's Llama-3 license. Dolphin was trained on data generated from GPT4, among other models.
See axolotl config
axolotl version: 0.4.0
base_model: meta-llama/Meta-Llama-3-8B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
tokenizerusefast: false
loadin8bit: false
loadin4bit: false
strict: false
model_config:
datasets:
- path: /workspace/datasets/dolphin-2.9/dolphin201-sharegpt2.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/Ultrachat200kunfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/dolphin-coder-translate-sharegpt2.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/dolphin-coder-codegen-sharegpt2.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/m-a-p_Code-Feedback-sharegpt-unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/m-a-p_CodeFeedback-Filtered-Instruction-sharegpt-unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/notsamanthanorefusals.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/Orca-Math-resort-unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/agentinstructreact_unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/toolbenchinstructj1s13kunfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/toolbenchnegativeunfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/toolbenchreact10p_unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/toolbenchtflancot30punfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/openhermes200k_unfiltered.jsonl
type: sharegpt
conversation: chatml
- path: /workspace/datasets/dolphin-2.9/SystemConversations.jsonl
type: sharegpt
conversation: chatml
chat_template: chatml
datasetpreparedpath: /workspace/datasets/dolphin-2.9/thingy
valsetsize: 0.0002
output_dir: ./out
sequence_len: 4096
sample_packing: true
padtosequence_len: true
gradientaccumulationsteps: 4
microbatchsize: 3
num_epochs: 3
logging_steps: 1
optimizer: adamw_8bit
lr_scheduler: cosine
learning_rate: 2e-5
wandb_project: dolphin-2.9-mixtral-8x22b
wandb_watch:
wandbrunid:
wandblogmodel:
trainoninputs: false
groupbylength: false
bf16: auto
fp16:
tf32: false
gradient_checkpointing: true
gradientcheckpointingkwargs:
use_reentrant: false
earlystoppingpatience:
resumefromcheckpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
savesperepoch: 4
savetotallimit: 2
save_steps:
evalsperepoch: 4
evalsamplepacking: false
debug:
deepspeed: deepspeedconfigs/zero3bf16.json
weight_decay: 0.05
fsdp:
fsdp_config:
special_tokens:
eostoken: "<|imend|>"
padtoken: "<|endof_text|>"
tokens:
- "<|im_start|>"
- "<|im_end|>"
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learningrate: 2e-05
- trainbatchsize: 3
- evalbatchsize: 3
- seed: 42
- distributedtype: multi-GPU
- numdevices: 8
- gradientaccumulationsteps: 4
- totaltrainbatchsize: 96
- totalevalbatchsize: 24
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lrschedulertype: cosine
- lrschedulerwarmupsteps: 7
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.146 | 0.0005 | 1 | 1.1064 |
| 0.6962 | 0.2501 | 555 | 0.6636 |
| 0.6857 | 0.5001 | 1110 | 0.6503 |
| 0.6592 | 0.7502 | 1665 | 0.6419 |
| 0.6465 | 1.0002 | 2220 | 0.6317 |
| 0.5295 | 1.2395 | 2775 | 0.6408 |
| 0.5302 | 1.4895 | 3330 | 0.6351 |
| 0.5188 | 1.7396 | 3885 | 0.6227 |
| 0.521 | 1.9896 | 4440 | 0.6168 |
| 0.3968 | 2.2289 | 4995 | 0.6646 |
| 0.3776 | 2.4789 | 5550 | 0.6619 |
| 0.3983 | 2.7290 | 6105 | 0.6602 |
Framework versions
- Transformers 4.40.0
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1
π GGUF File List
| π Filename | π¦ Size | β‘ Download |
|---|---|---|
|
dolphin-2.9-llama3-8b.Q2_K.gguf
LFS
Q2
|
2.96 GB | Download |
|
dolphin-2.9-llama3-8b.Q3_K_L.gguf
LFS
Q3
|
4.03 GB | Download |
|
dolphin-2.9-llama3-8b.Q3_K_M.gguf
LFS
Q3
|
3.74 GB | Download |
|
dolphin-2.9-llama3-8b.Q3_K_S.gguf
LFS
Q3
|
3.41 GB | Download |
|
dolphin-2.9-llama3-8b.Q4_0.gguf
Recommended
LFS
Q4
|
4.34 GB | Download |
|
dolphin-2.9-llama3-8b.Q4_1.gguf
LFS
Q4
|
4.78 GB | Download |
|
dolphin-2.9-llama3-8b.Q4_K_M.gguf
LFS
Q4
|
4.58 GB | Download |
|
dolphin-2.9-llama3-8b.Q4_K_S.gguf
LFS
Q4
|
4.37 GB | Download |
|
dolphin-2.9-llama3-8b.Q5_0.gguf
LFS
Q5
|
5.22 GB | Download |
|
dolphin-2.9-llama3-8b.Q5_1.gguf
LFS
Q5
|
5.65 GB | Download |
|
dolphin-2.9-llama3-8b.Q5_K_M.gguf
LFS
Q5
|
5.34 GB | Download |
|
dolphin-2.9-llama3-8b.Q5_K_S.gguf
LFS
Q5
|
5.22 GB | Download |
|
dolphin-2.9-llama3-8b.Q6_K.gguf
LFS
Q6
|
6.14 GB | Download |
|
dolphin-2.9-llama3-8b.Q8_0.gguf
LFS
Q8
|
7.95 GB | Download |