Josephgflowers/Qllama-5B-RAG-1
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: Josephgflowers/Qllama-5B-RAG-1
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-09-14 06:22:12 +08:00
commit 0150aae567
17 changed files with 624 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
.gitattributes vendored Normal file
View File

@@ -0,0 +1,51 @@
*.7z filter=lfs diff=lfs merge=lfs -text
*.arrow filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.bin.* filter=lfs diff=lfs merge=lfs -text
*.bz2 filter=lfs diff=lfs merge=lfs -text
*.ftz filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.h5 filter=lfs diff=lfs merge=lfs -text
*.joblib filter=lfs diff=lfs merge=lfs -text
*.lfs.* filter=lfs diff=lfs merge=lfs -text
*.model filter=lfs diff=lfs merge=lfs -text
*.msgpack filter=lfs diff=lfs merge=lfs -text
*.onnx filter=lfs diff=lfs merge=lfs -text
*.ot filter=lfs diff=lfs merge=lfs -text
*.parquet filter=lfs diff=lfs merge=lfs -text
*.pb filter=lfs diff=lfs merge=lfs -text
*.pt filter=lfs diff=lfs merge=lfs -text
*.pth filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
*.tar.* filter=lfs diff=lfs merge=lfs -text
*.tflite filter=lfs diff=lfs merge=lfs -text
*.tgz filter=lfs diff=lfs merge=lfs -text
*.xz filter=lfs diff=lfs merge=lfs -text
*.zip filter=lfs diff=lfs merge=lfs -text
*.zstandard filter=lfs diff=lfs merge=lfs -text
*.tfevents* filter=lfs diff=lfs merge=lfs -text
*.db* filter=lfs diff=lfs merge=lfs -text
*.ark* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*data* filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.meta filter=lfs diff=lfs merge=lfs -text
**/*ckpt*.index filter=lfs diff=lfs merge=lfs -text
*.safetensors filter=lfs diff=lfs merge=lfs -text
*.ckpt filter=lfs diff=lfs merge=lfs -text
*.gguf* filter=lfs diff=lfs merge=lfs -text
*.ggml filter=lfs diff=lfs merge=lfs -text
*.llamafile* filter=lfs diff=lfs merge=lfs -text
*.pt2 filter=lfs diff=lfs merge=lfs -text
*.mlmodel filter=lfs diff=lfs merge=lfs -text
*.npy filter=lfs diff=lfs merge=lfs -text
*.npz filter=lfs diff=lfs merge=lfs -text
*.pickle filter=lfs diff=lfs merge=lfs -text
*.pkl filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.wasm filter=lfs diff=lfs merge=lfs -text
*.zst filter=lfs diff=lfs merge=lfs -text
*tfevents* filter=lfs diff=lfs merge=lfs -text
merges.txt filter=lfs diff=lfs merge=lfs -text
tokenizer.json filter=lfs diff=lfs merge=lfs -text
vocab.json filter=lfs diff=lfs merge=lfs -text

30
README.md Normal file
View File

@@ -0,0 +1,30 @@
---
license: mit
---
This model needs further fine tuning.
See:
txtai-rag.py for rag implimentation with txt ai wikipedia.
Llamafyd version of Qwen .5B further fine tuned 1 epoch on wiki, math, science, and chat datasets. Based on Cinder data. This model further fine tuned 1 epoch on rag data.
Rough list of final datasets:
formatted_beaugogh-openorca-multiplechoice-10k.txt
formatted_BYC-Sophie-samsum-chatgpt-summary.txt
formatted_conversation_bio.txt
formatted_conversation_create_cinder_1.txt
formatted_conversation_Electrical-engineering.txt
formatted_conversation_multiturn_stem.txt
formatted_conversation_physics.txt
formatted_conversation_qa_rag_chem_prog_dataset.txt
formatted_conversation_qa_robot_ai_dataset.txt
formatted_conversation_qa_shopify_dataset1_rag.txt
formatted_conversation_qa_shopify_dataset_rag.txt
formatted_dyumat-databricks-dolly-5k-rag-split.txt
formatted_Hypoxiic-wikipedia-summary-subset1k-summary_token.txt
formatted_neural-bridge-rag-dataset-12000.txt
formatted_rachid16-rag_finetuning_data.txt
formatted_tiny_stories_1_summary_token_tag_token-xaa.txt
formatted_tiny_stories_2_summary_token_assistant-xah.txt
med_rag_small.txt
z_formatted_cinder_test.txt

32
config.json Normal file
View File

@@ -0,0 +1,32 @@
{
"_name_or_path": "/home/joe/Downloads/RAG-Summay/RAG-Summary",
"architectures": [
"LlamaForCausalLM"
],
"attention_bias": true,
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 1024,
"initializer_range": 0.02,
"intermediate_size": 2816,
"max_position_embeddings": 4096,
"max_window_layers": 21,
"mlp_bias": false,
"model_type": "llama",
"num_attention_heads": 16,
"num_hidden_layers": 24,
"num_key_value_heads": 16,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 1000000.0,
"sliding_window": 32768,
"tie_word_embeddings": true,
"torch_dtype": "float16",
"transformers_version": "4.41.0.dev0",
"use_cache": true,
"use_sliding_window": false,
"vocab_size": 151936
}

1
configuration.json Normal file
View File

@@ -0,0 +1 @@
{"framework": "pytorch", "task": "text-generation", "allow_remote": true}

11
generation_config.json Normal file
View File

@@ -0,0 +1,11 @@
{
"bos_token_id": 151643,
"do_sample": true,
"eos_token_id": [
151645,
151643
],
"repetition_penalty": 1.1,
"top_p": 0.8,
"transformers_version": "4.41.0.dev0"
}

BIN
merges.txt (Stored with Git LFS) Normal file
View File

Binary file not shown.

3
model.safetensors Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:ff655c413b73d2c107197a7e975300f24bcd648419dea17f25ec5fb22a15c7e0
size 928059544

3
optimizer.pt Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:2dd3a4a175c2ba2bd3e951f33ef04530b45254edf27496327f1f966e3dfb941c
size 3324282

3
rng_state.pth Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:386fcc8cc1089aade9450d86fb239ea3483f455fd2d78d8378645feecfec9d69
size 14244

3
scheduler.pt Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:52145e9f4901b4541897486decac06eff0f2707a063183b7c45ddf68fa72c4c2
size 1064

20
special_tokens_map.json Normal file
View File

@@ -0,0 +1,20 @@
{
"additional_special_tokens": [
"<s>",
"</s>"
],
"eos_token": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false
},
"pad_token": {
"content": "<|endoftext|>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false
}
}

3
tokenizer.json Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a12e3ba5d5e0ad173cf7b408ab8534c6be8cbc6a146714e9c7dc8cf2346603b1
size 7028043

43
tokenizer_config.json Normal file
View File

@@ -0,0 +1,43 @@
{
"add_prefix_space": false,
"added_tokens_decoder": {
"128247": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"128245": {
"content": "<s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"128247": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"<s>",
"</s>"
],
"bos_token": null,
"chat_template": "{% for message in messages %}{{'<s>' + message['role'] + '\n' + message['content'] + '</s>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% endif %}",
"clean_up_tokenization_spaces": false,
"eos_token": "</s>",
"errors": "replace",
"model_max_length": 32768,
"pad_token": "<|endoftext|>",
"split_special_tokens": false,
"tokenizer_class": "Qwen2Tokenizer",
"unk_token": null
}

285
trainer_state.json Normal file
View File

@@ -0,0 +1,285 @@
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 3643,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02744990392533626,
"grad_norm": 1.856189489364624,
"learning_rate": 4.865495470765852e-05,
"loss": 1.7566,
"step": 100
},
{
"epoch": 0.05489980785067252,
"grad_norm": 1.998369574546814,
"learning_rate": 4.728245951139171e-05,
"loss": 1.7366,
"step": 200
},
{
"epoch": 0.08234971177600878,
"grad_norm": 1.8302088975906372,
"learning_rate": 4.59099643151249e-05,
"loss": 1.7086,
"step": 300
},
{
"epoch": 0.10979961570134504,
"grad_norm": 1.9259424209594727,
"learning_rate": 4.453746911885809e-05,
"loss": 1.688,
"step": 400
},
{
"epoch": 0.1372495196266813,
"grad_norm": 1.6357027292251587,
"learning_rate": 4.316497392259127e-05,
"loss": 1.7272,
"step": 500
},
{
"epoch": 0.16469942355201755,
"grad_norm": 1.7630338668823242,
"learning_rate": 4.179247872632446e-05,
"loss": 1.7032,
"step": 600
},
{
"epoch": 0.19214932747735383,
"grad_norm": 2.0980162620544434,
"learning_rate": 4.041998353005765e-05,
"loss": 1.7523,
"step": 700
},
{
"epoch": 0.21959923140269008,
"grad_norm": 1.4029771089553833,
"learning_rate": 3.904748833379083e-05,
"loss": 1.6537,
"step": 800
},
{
"epoch": 0.24704913532802636,
"grad_norm": 1.5694669485092163,
"learning_rate": 3.767499313752402e-05,
"loss": 1.7099,
"step": 900
},
{
"epoch": 0.2744990392533626,
"grad_norm": 1.5211464166641235,
"learning_rate": 3.630249794125721e-05,
"loss": 1.688,
"step": 1000
},
{
"epoch": 0.30194894317869886,
"grad_norm": 1.527065396308899,
"learning_rate": 3.493000274499039e-05,
"loss": 1.6995,
"step": 1100
},
{
"epoch": 0.3293988471040351,
"grad_norm": 1.5144612789154053,
"learning_rate": 3.355750754872358e-05,
"loss": 1.6237,
"step": 1200
},
{
"epoch": 0.3568487510293714,
"grad_norm": 1.7216092348098755,
"learning_rate": 3.218501235245677e-05,
"loss": 1.6544,
"step": 1300
},
{
"epoch": 0.38429865495470766,
"grad_norm": 1.5754574537277222,
"learning_rate": 3.081251715618996e-05,
"loss": 1.6567,
"step": 1400
},
{
"epoch": 0.4117485588800439,
"grad_norm": 1.5684128999710083,
"learning_rate": 2.944002195992314e-05,
"loss": 1.6806,
"step": 1500
},
{
"epoch": 0.43919846280538016,
"grad_norm": 1.3534952402114868,
"learning_rate": 2.8067526763656326e-05,
"loss": 1.5953,
"step": 1600
},
{
"epoch": 0.46664836673071647,
"grad_norm": 1.4534939527511597,
"learning_rate": 2.669503156738952e-05,
"loss": 1.6146,
"step": 1700
},
{
"epoch": 0.4940982706560527,
"grad_norm": 1.3850529193878174,
"learning_rate": 2.53225363711227e-05,
"loss": 1.5718,
"step": 1800
},
{
"epoch": 0.5215481745813889,
"grad_norm": 1.536191701889038,
"learning_rate": 2.395004117485589e-05,
"loss": 1.6358,
"step": 1900
},
{
"epoch": 0.5489980785067252,
"grad_norm": 1.4949287176132202,
"learning_rate": 2.2577545978589075e-05,
"loss": 1.6173,
"step": 2000
},
{
"epoch": 0.5764479824320615,
"grad_norm": 1.489922046661377,
"learning_rate": 2.1205050782322264e-05,
"loss": 1.6291,
"step": 2100
},
{
"epoch": 0.6038978863573977,
"grad_norm": 2.090135097503662,
"learning_rate": 1.983255558605545e-05,
"loss": 1.6579,
"step": 2200
},
{
"epoch": 0.631347790282734,
"grad_norm": 1.798261284828186,
"learning_rate": 1.8460060389788635e-05,
"loss": 1.7008,
"step": 2300
},
{
"epoch": 0.6587976942080702,
"grad_norm": 1.43966805934906,
"learning_rate": 1.7087565193521824e-05,
"loss": 1.5593,
"step": 2400
},
{
"epoch": 0.6862475981334065,
"grad_norm": 1.5113532543182373,
"learning_rate": 1.571506999725501e-05,
"loss": 1.576,
"step": 2500
},
{
"epoch": 0.7136975020587428,
"grad_norm": 1.5978457927703857,
"learning_rate": 1.4342574800988198e-05,
"loss": 1.6075,
"step": 2600
},
{
"epoch": 0.741147405984079,
"grad_norm": 1.7422983646392822,
"learning_rate": 1.2970079604721382e-05,
"loss": 1.634,
"step": 2700
},
{
"epoch": 0.7685973099094153,
"grad_norm": 1.7618128061294556,
"learning_rate": 1.1597584408454571e-05,
"loss": 1.6333,
"step": 2800
},
{
"epoch": 0.7960472138347516,
"grad_norm": 1.4939252138137817,
"learning_rate": 1.0225089212187758e-05,
"loss": 1.5678,
"step": 2900
},
{
"epoch": 0.8234971177600878,
"grad_norm": 1.3491543531417847,
"learning_rate": 8.852594015920946e-06,
"loss": 1.5577,
"step": 3000
},
{
"epoch": 0.8509470216854241,
"grad_norm": 1.800729751586914,
"learning_rate": 7.480098819654131e-06,
"loss": 1.5562,
"step": 3100
},
{
"epoch": 0.8783969256107603,
"grad_norm": 1.4192384481430054,
"learning_rate": 6.107603623387318e-06,
"loss": 1.6218,
"step": 3200
},
{
"epoch": 0.9058468295360966,
"grad_norm": 1.487561583518982,
"learning_rate": 4.735108427120506e-06,
"loss": 1.6313,
"step": 3300
},
{
"epoch": 0.9332967334614329,
"grad_norm": 1.393309235572815,
"learning_rate": 3.3626132308536925e-06,
"loss": 1.5928,
"step": 3400
},
{
"epoch": 0.9607466373867691,
"grad_norm": 1.397757649421692,
"learning_rate": 1.990118034586879e-06,
"loss": 1.568,
"step": 3500
},
{
"epoch": 0.9881965413121054,
"grad_norm": 1.2685039043426514,
"learning_rate": 6.176228383200659e-07,
"loss": 1.5983,
"step": 3600
}
],
"logging_steps": 100,
"max_steps": 3643,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 3643,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 1.1045534257958093e+17,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}

3
training_args.bin Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:a4a252b43f42ad4d12be07cca1972ae8f64ffeeb2711dd370328e2c07e87a23c
size 5048

127
txtai-rag.py Normal file
View File

@@ -0,0 +1,127 @@
import os
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
import torch.nn as nn
from torchsummary import summary
from accelerate import dispatch_model, infer_auto_device_map
from txtai import Embeddings
from txtai.pipeline import LLM
#pip3 install git+https://github.com/neuml/txtai#egg=txtai[pipeline-llm]
# Wikipedia Embeddings Database
embeddings = Embeddings()
embeddings.load(provider="huggingface-hub", container="neuml/txtai-wikipedia")
#os.environ['OMP_NUM_THREADS'] = '6'
# Load model and tokenizer
model_path = "Josephgflowers/Qllama-.5B-RAG-1"#
# Define the device (CPU or GPU)
device = torch.device("cuda")
#device = torch.device("cpu")
model = AutoModelForCausalLM.from_pretrained(model_path,ignore_mismatched_sizes=True,torch_dtype=torch.float16).to(device)
print(model)
total_params = sum(p.numel() for p in model.parameters())
print("Total number of parameters: ", total_params)
sequence_length = 4096 # or whatever your specific sequence length is
embedding_size = 2048 # as per your model's definition
tokenizer = AutoTokenizer.from_pretrained(model_path)
stop_token = 128247
#'</s>' 2
def chat_with_model(prompt_text, stop_token, model, tokenizer):
# Encode the prompt text
encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False, return_tensors="pt").to(device)
# Generate response
output_sequences = model.generate(
input_ids=encoded_prompt,
#max_length=len(encoded_prompt[0]) + 256,
max_new_tokens=1024,
temperature=0.1,
repetition_penalty=1.2,
top_k=20,
top_p=0.9,
do_sample=True,
num_return_sequences=1,
eos_token_id=stop_token
)
# Decode the generated sequence
generated_sequence = output_sequences[0].tolist()
text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
response_text = text[len(prompt_text):].strip() # Extract only the response text
#response_text = response_text.replace("<s>","").replace("</s>","")
return response_text
# Initialize conversation history
conversation_history = ''
# Get user's preference for input mode and character name
input_mode = 'text' ##input("Enter 'text' for text input or 'speech' for speech input: ").lower()
character_name = '<|user|>' # input("Enter your character name (USER, JONAH, JOSEPH, KIMBERLY, etc.): ")
# Chat loop
num_chat = 1
while num_chat <= 20:
if input_mode == 'text':
question = input(f"{character_name}: ")
user_input = question # Get text input from user
context = "\n".join([x["text"] for x in embeddings.search(question)])
elif input_mode == 'speech':
print("Press Enter and then speak your input.")
input() # Wait for user to hit Enter
with sr.Microphone() as source:
print("Listening...")
audio_data = recognizer.listen(source, timeout=5, phrase_time_limit=10)
try:
user_input = recognizer.recognize_google(audio_data)
print(f"USER: {user_input}")
except sr.RequestError:
print("API request failed. Please try again.")
continue
except sr.UnknownValueError:
print("Could not understand the audio. Please try again.")
continue
else:
print("Invalid input mode. Please restart and choose a valid mode.")
break
#print('History: '+ conversation_history)
prompt_text = f"""
<s>
<|system|>
You will be given text as context to answer a users question.
</s>
<|data|>
{context}
</s>
<|user|>
{question}
</s>
<|assistant|>
"""
response_text = chat_with_model(prompt_text, stop_token, model, tokenizer)
response_text = response_text.replace('<s>','')
# Extract assistant's response from the response_text
response_text = response_text.split('</s>\n', 1)[0] # Extract the first message from the assistant
print(f"Assistant: {response_text}","\n")
# Update conversation history
conversation_history += f"<s>\n<|system|>\nYou will are a helpful assistant.</s>\n{character_name}\n{user_input}\n</s>\n<|assistant|>\n {response_text}</s>\n\n"
if len(conversation_history) > 3092:
conversation_history = conversation_history[2048:]
else:
conversation_history = conversation_history
num_chat += 1
#"You will receive a user question and context that should be used to answer the question."

BIN
vocab.json (Stored with Git LFS) Normal file
View File

Binary file not shown.