Update README.md

README.md

@@ -6,8 +6,8 @@ language:
 # DCLM-Edu
 
 ## Description
-This is a filtered version of the [DCLM](https://huggingface.co/datasets/mlfoundations/dclm-baseline-1.0) dataset using FineWeb-Edu educational quality [classifier](https://huggingface.co/HuggingFaceFW/fineweb-edu-classifier). We annotate each web page based on the educational quality
-on a scale from 0 to 5 and only keep samples with a score higher than 2. This dataset is intended for language models training and was used to train [SmolLM2-135M](https://huggingface.co/HuggingFaceTB/SmolLM2-135M) and [SmolLM2-360M](https://huggingface.co/HuggingFaceTB/SmolLM2-360M).
+This is a filtered version of [DCLM](https://huggingface.co/datasets/mlfoundations/dclm-baseline-1.0) dataset using FineWeb-Edu educational quality [classifier](https://huggingface.co/HuggingFaceFW/fineweb-edu-classifier). We annotate each web page based on the educational quality
+on a scale from 0 to 5 and only keep samples with a score higher than 2. This dataset is intended for small language models training and was used to train [SmolLM2-135M](https://huggingface.co/HuggingFaceTB/SmolLM2-135M) and [SmolLM2-360M](https://huggingface.co/HuggingFaceTB/SmolLM2-360M).
 
 **_Note:_** As show in the performance section, we find that further filtering the dataset to only keep **samples with `edu_int_score>=3` yields even better downstream performance when training small laguage models**. We include score 2 samples to allow for rebalancing and added diversity, but you can filter the dataset with `datasets` or `datatrove` as shown below.
 
@@ -53,20 +53,19 @@ pipeline_exec.run()
 
 ## Performance
 **Results of 360M ablation**
-We train a 360M model (using SmolLM2 setup) on 200B tokens from different datasets and evaluate on different benchmarks. DCLM-Edu denotes DCLM samples with an educational score higher than 3.
+We train a 360M model (using [SmolLM2](https://huggingface.co/HuggingFaceTB/SmolLM2-360M) setup) on 200B tokens from DCLM, FineWeb-Edu and DCLM-Edu and evaluate on different benchmarks. DCLM-Edu denotes DCLM samples with an educational score higher than 3.
+We find that the model trained on DCLM-Edu performs better on knowledge and reasoning tasks (MMLU & ARC):
 
-<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/Nt4tu5s4iaoN7ZqmqndJK.png" width="600" alt="image">
+<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/hOFJRusg6fEEtCpN-RJaP.png)" width="700" alt="image">
 
-DCLM-Edu gives consistent gains. The plot below shows the per-benchmark performance:
-<img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/L9sdCmfDVipTwDX5_dNcm.png" width="700" alt="image">
 
 We invite users to experiment with different data mixing depending on their model size.
 
 **Results of 1.7B ablation:**
-We also conducted some ablations at 1.7B scale but the gains for introducing the dataset mid-training durong SmolLM2 1.7B training weren't consistent with the ablation findings, so we only use the dataset for SmolLM2 135M and 360M. 
-We use an intermediate checkpoint of SmolLM2 1.7B (3T tokens) anddodoing a decay on different subsets of DCLM using the edu filtering:
+We also conducted some ablations at 1.7B scale, we use an intermediate checkpoint of SmolLM2 1.7B (3T tokens) and doing a decay on different subsets of DCLM using the edu filtering with thresholds 2, 3 and 4. 
 
 <img src="https://cdn-uploads.huggingface.co/production/uploads/61c141342aac764ce1654e43/ImwiEe712SN5TalxFOeeJ.png" width="700" alt="image">
+However we find that the gains from introducing this dataset mid-training during SmolLM2 1.7B training (which was trained on a mix of DCLM and FineWeb-Edu for 6T+ tokens) weren't consistent with the ablation findings, so we only use the dataset for SmolLM2 135M and 360M. 
 
 ## License
 Following DCLM-Baseline, this dataset is licensed under CC-BY-4.0.