lvj
/

Qwen3-4B-parq-2b-weight-4b-embed-shared

@@ -1,199 +1,163 @@
 ---
 library_name: transformers
-tags: []
 ---
-# Model Card for Model ID
-<!-- Provide a quick summary of what the model is/does. -->
-## Model Details
-### Model Description
-<!-- Provide a longer summary of what this model is. -->
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-### Model Sources [optional]
-<!-- Provide the basic links for the model. -->
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-## Uses
-<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
-### Direct Use
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-[More Information Needed]
-### Downstream Use [optional]
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-[More Information Needed]
-### Out-of-Scope Use
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-[More Information Needed]
-## Bias, Risks, and Limitations
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-[More Information Needed]
-### Recommendations
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
-## How to Get Started with the Model
-Use the code below to get started with the model.
-[More Information Needed]
-## Training Details
-### Training Data
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
-[More Information Needed]
-### Training Procedure
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-#### Preprocessing [optional]
-[More Information Needed]
-#### Training Hyperparameters
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
-#### Speeds, Sizes, Times [optional]
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
-[More Information Needed]
-## Evaluation
-<!-- This section describes the evaluation protocols and provides the results. -->
-### Testing Data, Factors & Metrics
-#### Testing Data
-<!-- This should link to a Dataset Card if possible. -->
-[More Information Needed]
-#### Factors
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
-[More Information Needed]
-#### Metrics
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
-[More Information Needed]
-### Results
-[More Information Needed]
-#### Summary
-## Model Examination [optional]
-<!-- Relevant interpretability work for the model goes here -->
-[More Information Needed]
-## Environmental Impact
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
-## Technical Specifications [optional]
-### Model Architecture and Objective
-[More Information Needed]
-### Compute Infrastructure
-[More Information Needed]
-#### Hardware
-[More Information Needed]
-#### Software
-[More Information Needed]
-## Citation [optional]
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
-**BibTeX:**
-[More Information Needed]
-**APA:**
-[More Information Needed]
-## Glossary [optional]
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
-[More Information Needed]
-## More Information [optional]
-[More Information Needed]
-## Model Card Authors [optional]
-[More Information Needed]
-## Model Card Contact
-[More Information Needed]

 ---
 library_name: transformers
+tags:
+- torchao
+- qwen
+- qwen3
+- nlp
+- chat
+- conversational
+language:
+- en
+base_model:
+- Qwen/Qwen3-4B
+pipeline_tag: text-generation
 ---
+# Quantization Recipe
+Follow the setup instructions at [hf-scripts/README.md](https://gitlab.com/lisjin-group/hf-scripts/-/blob/main/README.md).
+## QAT Finetuning with PARQ
+Run `bash scripts/train.sh` to finetune the model on a grade school math subset of [allenai/tulu-3-sft-olmo-2-mixture-0225](/datasets/allenai/tulu-3-sft-olmo-2-mixture-0225). You may need to adjust the variables in the script depending on your GPU count.
+After training completes, the final quantized model will be saved locally at `SAVE_DIR`.
+## Inference-ready Model Conversion
+Note: to `push_to_hub` you need to run
+```sh
+pip install -U "huggingface_hub[cli]"
+huggingface-cli login
+```
+and use a token with write access, from https://huggingface.co/settings/tokens
+To get the quantized model, run the following from the root of hf-scripts/:
+```py
+import logging
+import os
+import torch
+from functools import partial
+from huggingface_hub import whoami, get_token
+from torchao.quantization import quantize_
+from torchao.quantization.quant_api import IntxUnpackedToInt8Tensor
+from torchao.prototype.parq.quant import (
+    StretchedUnifTorchaoQuantizer,
+    UnifTorchaoQuantizer,
+)
+from torchao.prototype.parq.quant.config_torchao import _attach_hf_quantization_config
+from torchao.prototype.parq.quant.config_torchao import _get_config_from_quantizer
+from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
+from qat_utils import split_param_groups
+set_seed(0)
+model_path = f"{SAVE_DIR}"
+model = AutoModelForCausalLM.from_pretrained(
+    model_path, device_map="auto", dtype="auto"
+)
+tokenizer = AutoTokenizer.from_pretrained(model_path)
+params_quant, _, params_embed = split_param_groups(
+    model, embed_pat="(lm_head|embed_tokens)", printf=logger.warning
+)
+# Configure linear weights
+quant_bits = 2
+quantizer = (
+    StretchedUnifTorchaoQuantizer(quant_bits)
+    if quant_bits < 4
+    else UnifTorchaoQuantizer()
+)
+weight_only = False
+device = params_quant[0].device
+linear_config = _get_config_from_quantizer(
+    quantizer, weight_only, device, quant_bits, group_size=None
+)
+# Configure embeddings
+quantizer = UnifTorchaoQuantizer()
+embed_config = _get_config_from_quantizer(
+    quantizer, weight_only, device, quant_bits=4, group_size=128
+)
+def _filter_fn(module, *args, param_set) -> bool:
+    for p in module.parameters(recurse=False):
+        if p.data_ptr() in param_set:
+            return True
+    return False
+def _update_config(model, module_to_config, filter_fns, configs, tie_word_embeddings):
+    for name, module in model.named_modules():
+        if not hasattr(module, "weight"):
+            continue
+        elif tie_word_embeddings and name == "lm_head":  # skip tied weight
+            continue
+        for i, filter_fn in enumerate(filter_fns):
+            if filter_fn(module):
+                module_to_config[name] = configs[i]
+# Attach HF quantization_config
+embed_filter_fn = partial(_filter_fn, param_set={p.data_ptr() for p in params_embed})
+module_to_config = {"_default": linear_config}
+filter_fns = [embed_filter_fn]
+configs = [embed_config]
+_attach_hf_quantization_config(model, filter_fns, configs, module_to_config)
+# Convert quantized tensor classes to `IntxUnpackedToInt8Tensor`
+linear_filter_fn = partial(_filter_fn, param_set={p.data_ptr() for p in params_quant})
+quantize_(model, linear_config, filter_fn=linear_filter_fn)
+setattr(model.model.embed_tokens, "bias", None)
+quantize_(model, embed_config, filter_fn=embed_filter_fn)
+# Push to hub
+token = get_token()
+username = whoami(token=token)["name"]
+model_name = os.path.basename(model_path)
+save_to = os.path.join(username, model_name)
+model.push_to_hub(save_to, safe_serialization=False, commit_message=commit_message)
+tokenizer.push_to_hub(save_to)
+# Manual testing
+prompt = "John writes 20 pages a day. How long will it take him to write 3 books that are 400 pages each?"
+messages = [
+    {"role": "system", "content": "You are a helpful assistant."},
+    {"role": "user", "content": prompt},
+]
+templated_prompt = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True,
+)
+print("Prompt:", prompt)
+print("Templated prompt:", templated_prompt)
+inputs = tokenizer(templated_prompt, return_tensors="pt").to(model.device)
+inputs.pop("token_type_ids", None)
+start_idx = len(inputs.input_ids[0])
+response_ids = model.generate(**inputs, max_new_tokens=256, **kwargs)[0]
+response_ids = response_ids[start_idx:].tolist()
+output_text = tokenizer.decode(response_ids, skip_special_tokens=True)
+print("Response:", output_text)
+```
+The response from manual testing is:
+```txt
+To solve this problem, we need to determine how many days it will take John to write a total of 3 * 400 = 1200 pages.
+Given that John writes 20 pages a day, we can calculate the number of days needed by dividing the total number of pages by the number of pages written per day:
+\[
+\frac{1200 \text{ pages}}{20 \text{ pages/day}} = 60 \text{ days}
+\]
+Therefore, it will take John 60 days to write the 3 books.
+\[
+\boxed{60}
+\]
+```