average tokens

README.md

@@ -101,16 +101,50 @@ CDB has quite a diverse distribution of languages, overall there is about 143 la
 ## Tokens
 
 
-| Month | Docs | Tokens |
+Using [xlm-roberta-base](https://huggingface.co/docs/transformers/model_doc/xlm-roberta) tokenizer to estimate the number of tokens by sampling:
+
+
+
+### Token-length distribution
+
+Computed over a sample of **2,164,574 texts** from CDB-Dec, Oct, Nov.
+
+| Statistic | Value   |
+|-----------|---------|
+| Min       | 3       |
+| Max       | 702,667 |
+| Mean      | 1,354   |
+
+### Truncation rate by sequence length
+
+Percentage of texts that would lose tokens at each candidate `seq_len`.
+
+| seq_len | Truncation rate | Texts truncated |
+|---------|-----------------|-----------------|
+| 64      | 97.24%          | 2,104,865       |
+| 128     | 93.36%          | 2,020,754       |
+| 256     | 83.54%          | 1,808,331       |
+| 384     | 73.55%          | 1,591,996       |
+| 512     | 64.45%          | 1,394,971       |
+| 1024    | 37.96%          | 821,725         |
+
+**Note:** 1,394,971 texts (64.4%) exceed XLM-R's 512 max and will be truncated regardless of the `seq_len` chosen.
+
+
+On average we will have 
+
+
+| CDB | Docs | Tokens |
 |-------|------|--------|
-| Dec   |  XB    |   xT     |
+| Dec   |      |        |
 | Nov   |      |        |
 | Oct   |      |        |
 | **Total** | | |
 
+## Text-Attributed Graph w/ Cleaning
 
 
-## Text-Attributed Graph w/ Cleaning
+Overall across CDB-Dec, CDB-Nov, CDB-Oct